Full text of "Bulletin of the British Museum (Natural History) Zoo Supp"

9 * BS t> -^BULLETIN OF

2 2 FEB 1956

THE BRITISH MUSEUM
(NATURAL HISTORY)

ZOOLOGY
VOL II

1953-1955

PRINTED BY ORDER OF THE TRUSTEES OF

THE BRITISH MUSEUM (NATURAL HISTORY)

London: i$ss

DATES OF PUBLICATION OF THE PARTS

No.

17 February

1953

No.

5 March

1953

No.

7 November

1953

No.

22 January

1954

No.

22 January

1954

No.

24 September

1954

No.

22 October

1954

No.

22 October

1954

No.

31 December

1954

No.

10.

30 November

1954

No.

11.

25 January

1955

No.

12.

24 June

1955

No.

13-

19 August

1955

PRINTED IN

GREAT BRITAIN

AT THE

BARTHOLOMEW PRESS

DORKING

ADLARD AND SON, LTD.

CONTENTS

ZOOLOGY VOLUME 2

No. I. The Ostracod genus Trachyleberis. By J. P. Harding and P. C.

Sylvester-Bradley (Pis. 1-2) 1

No. 2. On the Puerulus stage of some spiny lobsters (Palinuridae) . By

Isabella Gordon 17

No. 3. On Sirpus, a genus of pigmy cancroid crabs. By Isabella Gordon 43

No. 4. The Macrotritopus problem. By w. j. rees (PL 3) 67

No. 5. Hearing in Cetaceans. By f. c. fraser and p. e. purves (Pis. 4-5) 101

No. 6. The "Rosaura" Expedition 1937-1938. (Pis. 6-9)

Gear, Narrative and Station List. By j. s. colman 119

Under- Water illumination and ecology in tropical estuaries. By

j. s. colman and l. h. n. cooper 131

The Echinodermata. By d. d. john and A. m. Clark 139

Fishes. Part 1 : Families Carcharhiniidae, Torpedinidae, Rosauridae,
(nov.), Salmonidae, Alepocephalidae, Searsidea, Clupeidae. By
denys w. tucker 163

Sponges. By m. burton 215

No. 7. On the Polyzoan genus Crepidacantha Levinsen. By d. a. brown 241

No. 8. Subdivisions of the genus Ptilinopus (Aves, Columbae). By A. j. cain 265

No. 9. A collection of Mesostigmatid mites from Alaska. By G. owen evans 285

No. 10. The Otariid seals of the Pacific Coast of America. By judith e. king

(Pis. 10-11) 309

No. 11. The Southern Right Whale Dolphin, Lissodelphis peroni (Lacepede).

By f. c. fraser (PL 12) 339

No. 12. New species of Tilapia (Pisces, Cichlidae) from Lake Jipe and the

Pangani River, East Africa. By rosemary h. lowe (Pis. 13-17) 347

No. 13. The morphology of the head of the hawfinch (Coccothraustes cocco-
thraustes) with special reference to the myology of the jaw. By

REGINALD WILLIAM SIMS 369

Index to Volume 2. 395

ERRATA

P. 65, first entry should read : Bohn, G. 1902 . . Bull. Sci. Fr. Belg., 36 : 178-551, etc.
P. 265 (title-page) : For " 2 Plates " read " 2 Figures."

1 2 MAR 1953

THE OSTRACOD GENUS
TRACHYLEBERIS

J. P. HARDING

AND

P. C. SYLVESTER-BRADLEY

BULLETIN OF

THE BRITISH MUSEUM (NATURAL HISTORY)

ZOOLOGY Vol. 2 No. i

LONDON: 1953

THE OSTRACOD GENUS TRACHYLEBERIS

J. P. HARDING v

British Museum (Natural History)

VV4

AND

P. C. SYLVESTER- BRADLEY .

Department of Geology, University of Sheffield

Pp. 1-16: Pis. 1-2; 25 Text-figures.

BULLETIN OF
THE BRITISH MUSEUM (NATURAL HISTORY)
ZOOLOGY Vol. 2 No. 1.

LONDON: 1953

THE BULLETIN OF THE BRITISH MUSEUM
(NATURAL HISTORY), instituted in 1949, is
issued in five series corresponding to the Departments
of the Museum, and an Historical Series.

Parts appear at irregular intervals as they become
ready. Volumes will contain about three or four
hundred pages, and will not necessarily be completed
within one calendar year.

This paper is Vol. 2, No. 1 of the Zoological series.

PRINTED BY ORDER OF THE TRUSTEES OF
THE BRITISH MUSEUM

Issued February, 1953. Price Four Shillings.

THE OSTRACOD GENUS TRACHYLEBERIS

By J. V. HARDING and P. C. SYLVESTER-BRADLEY

SYNOPSIS

When Brady erected the genus Trachyleberis for a species which he had described earlier as
Cy there scabrocuneata, he had before him specimens from New Zealand which did not in fact
belong to that species.

Both the type specimens of Cythere scabrocuneata and the specimens from New Zealand are
in the British Museum (Natural History) . A new name is proposed for the New Zealand species
and full descriptions of the type specimens of both species are given ; fortunately they are
congeneric.

Brady (1880) described the shell of a new species 01 ostracod, Cythere scabrocuneata,
from specimens collected by the " Challenger " Expedition. Later he (Brady, 1898)
described the limbs and other soft parts of further specimens from Lyttelton
Harbour, New Zealand, which he thought belonged to the same species, and as
the limbs showed some unusual features, he erected a new genus Trachyleberis.
Miiller (1912) and Skogsberg (1928) considered Trachyleberis to be synonymous
with Cythereis Jones, but, as one of us has shown elsewhere (Sylvester-Bradley,
19486), the shell of Trachyleberis differs significantly from that of Cythereis, and the
genus was therefore revived, though without reference to the soft parts.

Some of the original " Challenger " material is preserved in the British Museum
(Nat. Hist.) in the form of dry shells gummed to wooden microscope slides,
and the Lyttelton Harbour material, consisting of several specimens of both sexes,
is preserved in spirit in the Hancock Museum, Newcastle, and we are much indebted
to the Curator, Mr. C. E. Fisher, for letting us examine this material. He has also
graciously permitted some of the material to be kept at the British Museum. We
also re-examined the " Challenger " specimens and found that some of what Brady
refers to as " empty shells " had a few appendages still inside them ; and by treating
the whole ostracod with a 1% solution of tribasic sodium phosphate (Na 3 P0 4 ) it
has been possible to recover these appendages and make permanent mounts.

It is now quite evident that these " Challenger " specimens and the ones from
New Zealand are different species, and that the Lyttelton Harbour specimens must
be given a new name ; we propose to call them Trachyleberis lytteltonensis.

It is unfortunate that the only specimens in the British Museum (Nat. Hist.) of
the " Challenger " collections are from the Inland Sea of Japan, as Brady describes
C. scabrocuneata from three different localities, the Bass Straits, the Inland Sea of
Japan and Wellington Harbour, New Zealand, and it is quite possible that the
Wellington Harbour specimens were the same species as those from Lyttelton
Harbour, as both are from New Zealand harbours. Should these two species ever
be put into different genera the name Trachyleberis will have to remain with the
Japanese species, T. scabrocuneata, which must be based on the only available type

ZOOL. II, I.

4 THE OSTRACOD GENUS T RACHY LEBERI S

material. This is in spite of the fact that it was characters which Brady saw in
the New Zealand specimens, now called T. lytteltonensis , which led him to erect his
new genus.

The present paper attempts to describe these two species. Although it is intended
as a contribution to the taxonomy of the genus, we do not append a formal diagnosis.
We feel that as yet too few species are known by the details of both shell and soft
parts.

As the specimens from Lyttelton Harbour are more complete they will be described
first.

TRACHYLEBERIS Brady, 1898
Type-species (by monotypy) : Cythere scabrocuneata Brady, 1880.

Trachyleberis lytteltonensis sp. nov.
(Text-figs. 2-19 ; PI. 1, figs. 1-4, 7 ; PI. 2, figs. 1-4, 7, 8)

Trachyleberis scabrocuneata, Brady, 1898, Trans. Zool. Soc. Lond. 14 : 444, pi. 47, figs. i-y>

18-25.
Trachyleberis scabrocuneata, Hornibrook, 1952, New Zealand Geol. Surv., Palaeont. Bull. 18 :

32-33, pi. 3, figs. 38, 39, 48.

Material. About 40 specimens preserved in spirit from the Brady collection,
Hancock Museum, Newcastle-on-Tyne, dredged at 1-5 fathoms from Lyttelton
Harbour, New Zealand. The holotype and other figured specimens have been
mounted, and through the generosity of the Hancock Museum are now kept at
the British Museum (Nat. Hist.); these bear the following registered numbers:
1952.12.9.1-12.

Occurrence. In addition to its occurrence in the type locality T. lytteltonensis
has been reported by Hornibrook (1952, pp. 69, 71, under the name " T. scabro-
cuneata ") from four stations in the seas of the New Zealand area dredged at depths
ranging from 28-67 fathoms, and from one fossil locality in the Upper Miocene
(Tongaporutuan) of New Zealand.

Holotype. Male, B.M. 1952. 12. 9.1.

Paratypes. Both sexes, B.M. 1952 . 12 . 9 . 2-12.

Description. The carapace is subrectangular in lateral outline, with the dorsal
and ventral borders nearly straight, and converging but slightly towards the pos-
terior (PL 1, figs. 1-4). The anterior border is rounded, the posterior triangular,
the postero-dorsal margin being straight or slightly concave, the postero-ventral
margin curving evenly into the ventral margin. Sexual dimorphy is strong, the
males being longer in proportion than the females (see table below). In dorsal view
(PI. 2, figs. 1-4) the carapace is seen to be widest in the region of the sub-central
tubercle, and is compressed in the regions of the anterior and posterior plains.
In end view (Text-fig. 2) the carapace appears approximately triangular, both
valves showing a fairly abrupt ventral angulation, and the venter being almost

THE OSTRACOD GENUS T RA CHY LEB ERI S

flat. The left valve is larger than the right, its selvage fitting outside that of the
right throughout its course.

Dimensions: $ B.M. 1952. 12.9. 1
$ B.M. 1952. 12.9.6

Proportions: $ B.M. 1952. 12.9. 1
$ B.M. 1952. 12.9.6

eye

tubercle

hinge

tubercle

antero- dorsal
complex

anterior
plain

marqina

spines

anterior
marqina

antero-ventra!
complex

Length

1 -08 mm.
0-90 ,,

2-21

1-84
dorsal ridge

dorsal groove

Height

o • 49 mm.
0-49 „

1
1

posterior
hinge tubercle

Width

(of complete

carapace)

0-49 mm.

o-49 „
1 -oo
I -oo

postero- dorsal

complex

posterior
plain

sub-central
tubercle

ventral
ridges

posterior
marginal
spines

posterior
marginal

postero -ventral rim
complex

median
complex

Text-fig. i. Diagram to illustrate the terms used in the description of the

external ornament of the carapace in the Trachyleberidae.
Text-fig. 2. Trachyleberis lytteltonensis sp. nov. Profile in end view from the

posterior. B.M. 1952. 12.9.6.
Text-fig. 3. Trachyleberis lytteltonensis sp. nov. Internal lateral view of right

valve. B.M. 1952. 12.9.6.

6 THE OSTRACOD GENUS T RACHY LEBERI S

The following analysis of the ornament (see PL I, figs. 1-4) follows a scheme
previously put forward (Sylvester-Bradley, 1948a) illustrated by Text-fig. 1. Eye
tubercles are prominent on both valves, and run into the antero-dorsal complex,
which consists of a simple swelling. The sub-central tubercle forms a prominent
mammillated boss, and the antero-ventral complex consists of four tubercles.
Behind the sub-central tubercle the ornament is not clearly divided into ridges and
complexes, but consists of a number of tubercles, which have a tendency to bear
slightly mammillated summits (this character is not well shown by the photographs
reproduced in PL 1, which are of specimens coated with magnesia). There are
three ventral ridges, the uppermost bearing elongated tubercles. The anterior
marginal spines are divided into two groups, a dorsal set of smaller rounded spines,
and a ventral set of larger more pointed spines. The anterior marginal rim is
composed of a double line of tubercular spines, which supplement the armature
of the margin. The posterior marginal spines are prominent in the ventral region,
but almost absent dorsally. The posterior marginal rim is less well developed
than the anterior, and is armed with fewer tubercles. The anterior hinge tubercle
lies behind and above the eye tubercle in both valves, and the posterior hinge tubercle
of the left valve forms a prominent swelling near the commissure, best seen in dorsal
view.

The duplicature is fairly wide, extending 0-05 mm. inwards from the selvage in
anterior and postero-ventral regions, and up to 0-07 mm. at the posterior extremity.
The selvage is prominent in both valves, and follows a sinuous course along the ventral
margin, being markedly concave to the exterior in the centre of the venter. There
is no vestibule.

Radial pore canals (Text -fig. 3 ; PL 1, fig. 7) run to each marginal spine and to
each tubercle of the marginal rim. Some (presumably those serving the marginal
spines) terminate along the inner margin ; others (presumably those serving the
rim) terminate on the inner surface of the shell just inside the inner margin. The
total number of radial pore canals is therefore rather large, and they appear crowded,
and sometimes superimposed in lateral view. Most of them are inflated towards
the middle. The normal pore canals are few in number, one serving each tubercle
of the ornament in the same manner as the radial canals.

The sub-central tubercle appears as a muscle scar pit on the inside of the valve
(Text-fig. 3). The muscle scar pattern is that characteristic of so many of the
superfamily, with four slightly elongated scars, lying vertically above each other,
to the posterior of the pit, and a single scar, U-shaped, opening upwards,
within the pit. In addition there are two other scars, one above and one
below the main group. Each scar is raised as a tubercle on the inside of the
valve.

The antero-ventral complex forms a pit on the inner surface of the valve similar in
nature to the muscle-scar pit (Text-fig. 3).

The hinge of the left valve (PL 2, figs. 1-4, 7, 8) consists of two terminal slightly
elongated sockets, unbounded on the ventral side, and a median element subdivided
into an anterior rounded projecting tooth and an adjoining posterior finely crenulated
bar. In the right valve the anterior tooth projects beyond a less strongly projecting

THE OSTRACOD GENUS T RACHY LEBERI S 7

base. The posterior element is a slightly elongated lip-like tooth. The median
element is a groove subdivided as in the left valve to provide a deep anterior socket,
open ventrally, behind the anterior tooth.

In both valves a hollow lying immediately below and in front of the anterior
hinge element leads to the eye tubercle, and is for the reception of the eye of the
animal. The tubercle, being highly polished and transparent, perhaps serves as a
primitive lens.

The first antenna (Text-fig. 4) consists of six segments and is the same in both
sexes. The proportionate lengths of the segments and of the various setae, spines
and sense organs are best indicated by the drawing. The first segment has a tuft
of long spinules on the posterior face near the base. The second segment has a few
spinules on both posterior and anterior faces, and on the postero-distal corner
there is a slender, flexible, finely-haired seta. The third segment has a tapering
seta on the antero-distal corner. The fourth segment has two smooth setae on
the antero-distal corner, one stout and the other more slender, and from the middle
of the distal edge on the medial side is a smooth, slender seta. The fifth segment
has two stout setae and a slender one near the antero-distal corner, and one slender
one on the distal edge on the medial side. These four are all smooth. The last
segment has two stout terminal setae ; and a slender one which is united at its base
to a sense club (s.).

The second antenna (Text-figs. 5 and 6) consists of a protopod (p.) of one segment,
an endopod {en.) of three segments, and a very small exopod {ex.) of two segments,
and is alike in both sexes except for the small differences in the exopod. The
protopod is quite unarmed. The exopod is reduced to what appears to be a sense
organ. This is best developed in the female (Fig. 6), where it is very thinly chitinized
and rather broad, and is less than twice the length of the first segment of the endopod.
In the male (Text-fig. 5) the exopod is shorter still and looks very like an ordinary
seta. Brady (1898) describes the exopod as " a very short, falcate, urticating
seta [in the female], which is absent in the male."

The first segment of the endopod bears some stiff bristles and a long slender seta
with very minute hairs on the postero-distal corner. The second segment of the
endopod bears a tuft of hair-like spinules on its first quarter on the inner side.
About two-thirds along its length there are two smooth setae on its anterior face,
and on the posterior face at about this level there is one stout, tapering seta, and
one longer but more slender one, both feathered, and near them a sense club (s.).
At the postero-distal corner of this segment there is one stout, feathered seta and one
small naked one. The last segment has three stout, naked setae, two together
half-way along its posterior margin and the other one terminal.

The mandible (Text-fig. 8) is the same in both sexes. The shape of the basal
biting part is shown in the figure. The biting edge contains six strong teeth each
with a secondary cusp. The most anterior of these teeth is the largest, and they
decrease in size gradually towards the posterior corner ; the first two take up half
the biting edge. There is a bristle between the first and second and another between
the second and third teeth, and on the posterior medial corner beyond the largest
tooth there are two ribbon-like bifurcate setae which are about twice the length of

THE OSTRACOD GENUS T RACHY LEB ERI S

Text-figs. 4-12. Trachyleberis lytteltonensis sp. nov.

Fig. 4, first antenna, x 130. Fig. 5, second antenna of male, x 130. Fig. 6, part of
second antenna of female with exopod, x 180. Fig. 7, brush-shaped organ of male,
right member from behind, x 360. Fig. 8, right mandible from inner side, x 130.
Fig. 9, left maxilla from outer side, x 130. Fig. 10, palp and endites of left maxilla
from outer side, x 220. Fig. 11, posterior part of body of female, x 180. Fig. 12,
posterior part of body and copulatory apparatus of male, x 130. b., postero-ventral
corner of copulatory appendage ; ch., chamber ; cop., copulatory appendage ; en.,
endopod ; ex., exopod ; fl., flagellum ; fu., f ureal ramus ; ly., labyrinth ; od., ovi-
duct ; p., protopod ; pe., penis ; r.s., receptaculum seminis ; s., sense club ; /.,
terminal seta ; va., vagina ; v.d., vas deferens ; vr., verruca.

THE OSTRACOD GENUS T RACHY LEB ERI S g

the teeth near them. Beyond them again there is a single bent seta. Between the
biting edge and the palp there is a seta with a few hairs.

The palp consists of the second segment of the protopod bearing a three-segmented
endopod and an exopod of one segment. The exopod rises from the dorsal part
of the outer face of the protopod and is directed distally. It consists of a single
rounded segment narrow at its base, with five carrot-shaped, hairy setae which are
thinly chitinized and of very different lengths, as shown in the figure.

The ventral margin of the protopod carries two delicate setae. The one nearest
the distal corner is more or less parallel to the endopod and is shorter and stouter
than the other. Both are annulated and hairy.

The first segment of the endopod has an annulated hairy seta on the inside of the
distal dorsal corner, and on the ventral distal corner there are four setae, of which
the two medial ones are small. The two on the outer side are stout and curved
and are armed with long, stiff hairs which are rather wide apart. In preparations
these are most evident in the outer seta, in which the hairs lie parallel to the cover
slip. On the dorsal part of the distal end of the next segment is a group of seven
whip-like, flexible, naked setae, and on the ventral distal corner there is a long
tapering seta with hairs. On the medial side in the middle of the distal edge there
is a seta which is a little longer than the terminal segment and has a few stiff hairs
on the ventral side.

The terminal segment of the endopod is a little longer than the preceding one,
but rather less than half its diameter and slightly tapering. It has four slender
terminal setae, one of which has hairs.

The maxilla (Text-figs. 9, 10) is the same in both sexes. Between the long,
plumose setae of the vibratory plate and the palp on the distal edge of the appendage
there is a hairy swelling with a group of bristles directed parallel to the palp. The
first of the 16 plumose setae arises from this swelling. This seta is constricted into
two parts, both of which are rather swollen. The distal part tapers to a point.
The second to the fifteenth plumose setae are longer and very much alike. They
have a swollen basal part rather longer than the corresponding basal part of the
first seta, and the distal part is not swollen but long and feathered. The sixteenth
seta is much smaller than the others but hardly longer than the basal part of the
middle 14, and it appears to be of one piece without a constriction.

The palp is composed of two more or less cylindrical segments, the first segment
being about twice the length and twice the width of the second. On the distal
edge of the first segment and dorsal to the second segment are four slender setae
differing from each other in hairiness. That next to the second segment is the
shortest, being about twice the length of the second segment ; it has minute hairs
all round it. The other three setae are all nearly twice this length ; one appears
to be naked, the second has rather short hairs, and the other has long hairs ; these
three are annulated for the distal part of their length. In addition to these setae,
the distal margin of the first segment carries a seta near the outer face of the second
segment. This seta has a swollen base from which arises a short secondary setule.

There are three distinct endites, the first, i.e., the one nearest the palp, bears
terminally eight tapering setae, many of which have swollen bases. These are all

io THE OSTRACOD GENUS T RACHY LEBERI S

naked or nearly so. The two other endites have several pointed terminal setae,
the actual number being difficult to determine. In addition to these terminal
setae, the last endite has a fairly long seta arising from near its base on the side
facing the palp.

The fifth, sixth and seventh pairs of appendages are basically similar to one
another. They are all leg-like, and will be referred to here as the first, second and
third pairs of legs. Each of the three pairs shows sexual dimorphism, and further-
more, in the first pair of legs of the male the right leg differs from the left leg. The
following description will apply to both sides and both sexes unless otherwise
indicated.

The first leg (Text-figs. 13-15) consists of four segments. The first segment has
two annulated setae on the anterior margin and two on the antero-distal corner
overhanging the knee-like joint. Near the base of the leg on the posterior side
there is a soft, hairy, carrot-shaped seta, and near it along the posterior margin of
the segment there are some long hairs. The postero-distal corner of the first segment
has a hollow surrounded by fine spinules. The second segment broadens distally
and carries one seta. This is on the antero-distal corner in the female (Text -fig. 13),
and carries short hairs. In the male this seta is more hairy and placed on the
anterior margin short of the corner, and is farther from the corner on the right leg
(Text-fig. 14) than on the left (Text-fig. 15). Apart from a few minute spinules the
right leg of the male bears nothing else on the second segment, but on the left leg
of the male there is a swelling in about the middle of the anterior margin. This
carries a roughened friction pad on the medial side and a hairy boss on the outer
side. In the female (Text-fig. 13) there is a group of spinules on this part of the
anterior margin. The third and fourth segments of the leg are the same in both
sexes and on both sides. The distal edge of each segment is armed with small
spinules on both sides ; these are much longer laterally than medially. The terminal
curved claw is about twice the length of the last segment and is smooth in the male
and pectinate on the inside of the curve in the female.

The second pair of legs (Text-figs. 16, 17) is sexually dimorphic, but in the male
as well as in the female the left and right legs form a symmetrical pair. The first
segment is the same in both sexes, and has similar setae to those of the first leg
except that only one seta overhangs the " knee." The second segment broadens
gradually distally, and is nearly as long as the first segment but much narrower.
It bears an annulated seta on the antero-distal corner, and in the male (Text-fig. 17)
three groups of stiff, sharply-pointed bristles ; in the female (Text-fig. 16) only
the most proximal of these three groups is present. The third and fourth segments
are similar to those of the first leg but rather longer and more slenderly built. The
claw is smooth in the male and pectinate in the female.

The third pair of legs (Text-figs. 18, 19) is essentially similar to the second pair.
The first segment has a normal annulated seta on the postero-proximal corner
instead of the carrot-shaped seta. In the male the second segment has four groups
of conspicuous, straight, sharply-pointed bristles on the anterior margin ; only
the two most proximal of these groups are represented in the female, and in this
sex the bristles are more flexible. The seta on the antero-distal corner is long,

THE OSTRACOD GENUS T RACHY LEB ERI S n

and in the male, crooked. The male has a group of slender-pointed spines on the
middle of the anterior margin of the last segment. The terminal claw is pectinate
in both sexes.

The copulatory apparatus is of the usual complex nature. Specimens were
unfortunately not well enough preserved for the details of the softer parts to be
made out. The structure of the less easily macerated parts is shown in Text-fig. 12.
The following features were found in each of five males examined. On each side
the apparatus consists of two parts, a roughly oval muscular part called by Skogs-
berg (1928) and others the "penis" (pe.) and a distal . triangular "copulatory
appendage " (cop.). The muscular part contains a number of chitinous struts
with muscles running between them as shown. Near the base of the copulatory
appendage there is a very complex labyrinth (ly.) with a rounded chamber (ch.)
which has chitinous walls. A tube with spirally thickened walls, presumably the vas
deferens (v.d.), runs to this chamber. The other end of the tube was found in very
different positions in different specimens, having apparently become free as a result
of maceration of the soft parts ; but the position figured is believed to be correct.
The copulatory appendage is a thin-walled, roughly triangular organ with irregular
outgrowths from its broader distal part. These outgrowths are roughly similar
on both sides ; but those of the same side of two different individuals are usually
more alike than the right and left ones of the same individual. The appendage
on the right has a flagellum (fl.) which is absent on the left. The " postero-ventral"
part of the copulatory appendage ends with a rounded point (b.). We were unable
to determine a ductus ejaculatorius.

The female reproductive system and the shape of the posterior part of the body
are shown in Text-fig. 11. On each side of the body near the caudal furca, which
in this sex has three setae, lies the verruca (vr.). The oviduct (od.) is a sinuous
tube connected with the verruca. The vagina (va.) is separate from the oviduct
and opens anterior to it. Both these ducts are connected with the receptaculum
seminis (r.s.). The receptaculum and the other organs are all paired.

The paired brush-shaped organs of the male (Text -fig. 7) are placed on the ventral
side of the body between the " knees " of the first pair of legs.

The f ureal rami (/«.) are reduced to a simple swelling on each side bearing three
setae in the female and two setae in the male. In both sexes the hindermost part
of the body bears a single median seta (t.).

Trachyleberis scabrocuneata (Brady)
(Text-figs. 20-25 ; PI. 1, figs. 5, 6, 8 ; PL 2, figs. 5, 6, 9, 10.)

Cythere scabrocuneata Brady, 1880, Challenger Rep. 1 : 103, pi. 17, figs. 5a-f ; pi. 23, figs.

2a-c.
Not Trachyleberis scabrocuneata, Brady, 1898, Trans, zool. Soc. Lond. 14 : 444, pi. 47, figs.

1-7, 18-25.
Trachyleberis scabrocuneata, Sylvester-Bradley, 1948, J. Paleont. 22 : 794, pi. 122, figs. 13-18.
Not Trachyleberis scabrocuneata, Hornibrook, 1952, New Zealand Geol. Surv., Palaeont. Bull.

18 : 32-33, pi. 3, figs. 38, 39, 48.

12 THE OSTRACOD GENUS TRACHYLEBERIS

Occurrence and Material. The only material that has survived is that
dredged by the " Challenger " Expedition from the Inland Sea of Japan at 14
fathoms. The following specimens are held in the British Museum (Nat. Hist.) :

Lectotype (here chosen) B.M. 1952. 12. 10. 1, 2.

Paratypes : B.M. 1948 . 3 . 10 . 1-5, B.M. 1952 .12.10. 3-9.

Other Records. Brady's other records of this species are suspected to be
T. lytteltonensis or other species.

Text-figs. 13-19. Trachyleberis lytteltonensis sp. nov.

Fig. 13, left first leg of female, x 180.
15, left first leg of male, x 145. Fig.
left second leg of male, x 145. Fig.
left third leg of male, x 145.

Fig. 14, right first leg of male, x 145. Fig.
16, left second leg of female, x 180. Fig. 17,
18, left third leg of female, x 180. Fig. 19,

THE OSTRACOD GENUS T RACHY LEB ERI S 13

Description. The shell of this species has been previously described (Sylvester-
Bradley, 1948&) on the basis of the type material here re-examined, and only such
details as serve to distinguish the species from T. lytteltonensis need be dealt with
again. The shell of the species has been refigured on Pis. 1 and 2 to aid comparison.

There are no differences to report in the general shape of the carapace, the degree
of sexual dimorphy, or the size. In the following table of dimensions and propor-
tions, those of the complete carapaces have been estimated from single valves :

Length

Height

Width

Dimensions :
Proportions

£ B.M. 1948.3. 10. 1
? B.M. 1948.3. 10.5
6* B.M. 1948.3. 10. 1
$ B.M. 1948.3. 10.5

1 -io mm.
0-90 ,,

2-13

1-94

0-52 mm.
0-46 ,,

• 49 mm
0-46 ,,
0-94
1 -oo

The antero-dorsal complex of T. scabrocuneata (PL 1, figs. 5, 6, 8) consists of a
sharp ridge running from the eye tubercle forwards and downwards to as far as the
median line, and presents an immediate contrast to the rounded swelling which forms
the antero-dorsal complex of T. lytteltonensis. Similarly the sub-central tubercle
of T. scabrocuneata is surmounted by ridge-like elevations rather than the mammillae
of T. lytteltonensis. The general style of ornament is similar in the two species, but
the tubercles in T. scabrocuneata are smaller, more sharply defined and greater in
number than in T. lytteltonensis. In T. scabrocuneata the marginal spines in the
post ero- ventral region, forming a ventral ridge, are compressed and blade-like,
and swell out above the base (PI. 2, figs. 9 and 10). It is interesting to note that
a third species, T. thomsoni Hornibrook, 1952, ranging in New Zealand from the
Paleocene (Waipav/an) to Recent, as shown by Hornibrook's figures, is also similar
in style of ornament, but the tubercles are even fewer, and are more pronounced,
rising from dorsal and ventral ridges as veritable spines.

The internal characters of the carapace (PI. 2, figs. 5, 6, 9 and 10) are essentially
similar to those displayed by T. lytteltonensis. The anterior tooth of the right
valve, however, projects far more strongly above its base than that of T. lytteltonensis
(PL 2, fig. 6).

The first antenna of the female is shown in Text-fig. 20. This drawing has been
made with the aid of a camera lucida, but antennules from three separate females
have been used in order to include as many setae as possible. In none of the speci-
mens was the basal segment or the terminal seta with attached sense club present.
Lightly stippled " ghosts " of these have been added to the drawings by analogy
with the other species. As will be seen by comparing Text-fig. 20 with Text-fig. 4
the first antennae of the two species are practically identical. The second antennae
also appear to be identical in the two species (compare Text-fig. 21 with Text-fig. 5).
Only the three terminal segments of this appendage with the five most terminal
setae are preserved, however, and we could find no exopod.

The mandible (Text-fig. 22) is similar in the two species, though the palp has less
than seven whip-like setae on the posterior side of the penultimate segment, but this
may have been because some of the setae had been lost. The palps of two mandibles

THE OSTRACOD GENUS T RACHY LEB ERI S

were recovered, but unfortunately in both examples the first segment of the palp
was rather distorted and in neither was it possible to see the exopod.

A number of parts of the legs were preserved, but the only complete one is the

Text-figs. 20-25. Trachyleberis scabvocuneata Brady. Type material.

Fig. 20, first antenna of female, x 165, drawn from three specimens. Fig. 21, second
antenna, x 165, from two specimens. Fig. 22, mandible, x 165. Fig. 23, second
leg of female, x 260. Fig. 24, three segments of the third leg of a male, x 260.
Fig. 25, copulatory apparatus of male, x 260, drawn from three specimens.

Segments and setae shown lightly stippled were missing in the specimens
seen, and have been added by analogy with the other species.

second left leg of a female shown in Text-fig. 23. Instead of the group of bristles
present on the proximal third of the second segment as in T. lytteltonensis, there
are two groups of short spinules placed laterally. Otherwise there seems to be no
difference of any importance. Three segments of the third right leg of the male

THE OSTRACOD GENUS T RACHY LEB ERI S 15

are shown in Text-fig. 24. These are very similar to the corresponding segments
of T. lytteltonensis (Text-fig. 22), and the ventral margins of the second and third
segments are hairy. These hairs, however, do not seem to be arranged in distinct
groups as they are in T. lytteltonensis.

The copulatory apparatus (Text-fig. 25), of which no less than three specimens
were preserved, is quite unlike that of T. lytteltonensis (compare Text-fig. 25 with
Text-fig. 12). None of the three specimens was sufficiently well-preserved to show
the muscles, but all three were remarkably like one another ; two of the three
were right- and left-hand members of one specimen. The apparatus in the two
species is so different that it is difficult to homologize the parts. In T. scabrocuneata
there is a distinctly spiral ejaculatory duct which could not be seen in T. lytteltonensis.
The distal parts of the copulatory appendage in T. scabrocuneata is a long, rather
curved organ with heavily chitinized walls, as shown in the figure, and very different
from the broadly expanded, laminate and delicately-built organ of T. lytteltonensis.

REFERENCES

Brady, G. S. 1880. Report on the Ostracoda dredged by H.M.S. " Challenger " during the

years 1873-1876. Rep. Sci. Res. " Challenger " Zool. 1 : 1-184, pis. 1-44.
1898. On new or imperfectly known species of Ostracoda, chiefly from New Zealand.

Trans. Zool. Soc. Lond. 14 : 429-452, pis. 43-47.
Hornibrook, N. de B. 1952. Tertiary and Recent Marine Ostracoda of New Zealand.

Palaeont. Bull., N.Z., 18 : 1-82, pis. 1-18.
Muller, G. W. 191 2. Crustacea, Ostracoda. Das Tierreich, Lief. 31 : i-xxxiii, 1-434,

92 Text-figs.
Skogsberg, T. 1928. Studies on Marine Ostracods. Part II : External morphology of

the Genus Cythereis with descriptions of twenty-one new species. Occ. Pap. Calif. Acad.

Sci. 15 : 1-154, pis. 1-6, 23 Text-figs.
Sylvester-Bradley, P. C. 1948*3. Bathonian Ostracods from the Boueti Bed of Langton

Herring, Dorset. Geol. Mag. 85 : 185-204, pis. 12-15, 7 Text-figs.
19486. The Ostracode Genus Cythereis. J. Paleont. 22 : 792-797, pi. 122, 1 Text-fig.

EXPLANATION OF PLATES
PLATE 1

All external lateral views. All figures except fig. 7 from retouched photographs of specimens
coated with magnesium oxide, x 60. Fig. 7 from an unretouched photograph, x 120, of a
specimen immersed in benzyl alcohol and illuminated by transmitted light.

Figs. 1-4, 7. Trachyleberis lytteltonensis sp. nov. Figs. 1, 2, B.M. 1952. 12.9. 1 (<$) ;
fi gs. 3, 4, 7, B.M. 1952. 12.9.6 ($).

Figs. 5, 6, 8. Trachyleberis scabrocuneata (Brady). Fig. 5, B.M. 1948.3. 10. 1 (<$) ; fig.
6, B.M. 1948.3.10.2(0*); fig. 8, B.M. 1948.3. 10.5 ($).

PLATE 2

All figures are from retouched photographs, x 60. Figs. 1-6, dorsal views ; figs. 7-10,
internal lateral views.

Figs. 1-4, 7, 8. Trachyleberis lytteltonensis sp. nov. Figs. 1, 2 B.M. 1952. 12.9. 1 (<J) ; figs.
3, 4, 7, 8, B.M. 1952. 12.9.6 (?).

Figs. 5, 6, 9, 10. Trachyleberis scabrocuneata (Brady). Figs. 5, 9, B.M. 1948.3. 10. 1 (<$) ;

figs. 6, 10, B.M. 1948. 3. 10. 5 ($),

PRESENTED

X' L - 1 O MAD 1Qi;7

\T E

TRACHYLEBERIS

Hull. B.M. (N.H.) Zool. 11, i

PLAT E

TRACHYLEBERIS

PRINTED IN GREAT BRITAIN BY
AD LARD AND SON, LIMITED
BARTHOLOMEW PRESS, DORKING

1 2 MAR 1953

ON THE PUERULUS STAGE

OF SOME SPINY LOBSTERS

(PALINURIDAE)

ISABELLA GORDON

BULLETIN OF

THE BRITISH MUSEUM (NATURAL HISTORY)

ZOOLOGY Vol. 2 No. 2

LONDON: 1953

ON THE PUERULUS STAGE OF SOME SPINY
LOBSTERS (PALINURIDAE)

ISABELLA GORDON D.Sc, Ph.D.

British Museum (Nat! Hist.)

Pp. 17-42; 9 Text-figures

BULLETIN OF
THE BRITISH MUSEUM (NATURAL HISTORY)
ZOOLOGY Vol. 2 No. 2

LONDON: 1953

THE BULLETIN OF THE BRITISH MUSEUM
(NATURAL HISTORY), instituted in 1949, is
issued in five series corresponding to the Departments
of the Museum, and an Historical Series.

Parts appear at irregular intervals as they become
ready. Volumes will contain about three or four
hundred pages, and will not necessarily be completed
within one calendar year.

This paper is Vol. 2, No. 2 of the Zoological series.

PRINTED BY ORDER OF THE TRUSTEES OF
THE BRITISH MUSEUM

Issued March, 1953 Price Six Shillings

ON THE PUERULUS STAGE OF SOME SPINY
LOBSTERS (PALINURIDAE)

By ISABELLA GORDON

SYNOPSIS

The paper includes a key to the determination of the adult Palinurid genera to which the
known post-larvae are referable, descriptions and figures of the Puerulus stage of several species
of the genus Panulirus and notes on the Puerulus of the European Spiny Lobster. In an
appendix the range of variation in the adults of the species Panulirus homavus (L.) is discussed.

INTRODUCTION

In March, 1952, I received from Hugh Copley, Esq., Fish Warden, Nairobi, an
almost complete and beautifully preserved specimen of the Puerulus stage of a
Palinurid. The antennae were intact, and the most striking feature of the post-
larva was the dark brown spatulate apex of the long flagelmm. A very similar
Puerulus has recently been described in a Japanese journal which is not yet avail-
able in this country ; as the text is mostly in Japanese a description in English
seems essential. A number of Pueruli from the Marquesas and Coiba were found
in undetermined collections in the British Museum and they are also described
in the following short paper. All these post-larvae proved to belong to the genus
Panulirus, and it has been possible to refer most of them to their respective adult
species. All possess a pair of long, backwardly directed spines at the posterolateral
angles of the thoracic sternum (near the bases of the last pair of walking legs).
All agree, moreover, in having the exopodite of maxilliped 3 vestigial, and therefore
belong to species in which that exopodite is lacking in the adult. The material
from the Marquesas is referable to either Panulirus homarus (L.) or the closely
related P. dasypus (Latr.) and, in an appendix, I discuss my reasons for believing
that these two may belong to one variable species.

In the Puerulus referred to P. argus (Latr.) by Gurney (1942, p. 234, fig. 93) the
posterolateral thoracic spines are replaced by blunt nodules, and the exopodite
of maxilliped 3 is rather longer than the ischium of the endopodite. The postero-
lateral spines are absent in Puerulus pellucidus (Ortmann), which has been referred
to Panulirus japonicus (von Siebold) — see synonymy in Holthuis, 1946, p. in,
and Nakazawa, 1917.

At the present time the post-larva is known for only four genera of the Pali-
nuridae, namely, Palinurus Fabr., Panulirus White, Jasus Parker and Justitia
Holthuis. I have examined Pueruli belonging to each of the first three genera
but that of Justitia is known only from a single specimen collected by the " Blake."
It was first described as the " Puerulus d'Agassiz," referable to Palinurus longi-
manus H. M.-Edw. by Bouvier (1913, pp. 82 and 87 ; 1914, p. 187 ; and 1925,

ZOOL. II, 2. §

18 ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS

p. 442, pi. 8, fig. 1), and Bouvier's figure has been reproduced by Schmitt (1935,
p. 173, fig. 37). Holthuis erected the genus Justitia for/, longimana (H. M.-Edw.)
and its variety mauritiana (Miers) in 1946, p. 115.

In the British Museum collection are the three specimens — Phyllosoma, Phyllo-
soma with the Puerulus in the act of disengaging itself, and Puerulus — of Palinurus
vulgaris Latr. figured by Bouvier (1914, figs. 1, 2, and 3-6). Bouvier's sketch of
the carapace of the Puerulus is not very accurate as to the anterior spines, the
supraorbital spines especially being much more prominent than in the actual speci-
men (reg. no. 1914.12.17.4). I take this opportunity of refiguring the anterior
part of the carapace and the antennule (Fig. la, b and c) for comparison with those
oijasus (Fig. id and e) and of Justitia as figured in Bouvier, 1925, pi. 8, fig. 1 (which
appears to be accurate). Incidentally, Holthuis says that the common European
" langouste " should be known as Palinurus elephas (Fabr.), which antedates
vulgaris Latr.

I also give a key to the determination of the adult genera to which the post-
larvae are referable ; it is more detailed than that given by Bouvier (1913, p. 8y),
but for the characters of the Puerulus of Justitia I have had to rely on Bouvier's
figure and descriptions. I am not sure whether the antennae are intact, but if
so they do not exceed 1-5 times the body length.

Key to the Determination of the Adult Genera to which the
Post-larvae are Referable

I. Antennular flagella distinctly shorter than the peduncle, outer much more robust than
the inner (Fig. \a and d). Anterior margin of the antennular 1 tergum much narrower
than the space between the tips of the supraorbital spines. Rostrum well formed or
rudimentary (Fig. la and d). The " brevicornis " group.

a. Exopodite of maxilliped 3 long, with multi-segmented flagellum. Posterolateral
spines present on several of the somites of thoracic sternum. Rostrum rudi-
mentary.

1. Supraorbital spines denticulate on upper margin ; plate formed by
these spines and the rostrum broad and advanced so as to conceal
the eye-stalks, the corneae visible laterally. Three pairs of spinules
on thoracic sternum. Distinct median carina on carapace and
abdominal somites 2-5. Only a few spines on carapace near the
orbits. Second segment of antennular peduncle equal to third

Justitia Holthuis.
1 The antennal tergum of some authors.

Fig. 1. Palinurus elephas (Fabr.) = vulgaris Latr. The Puerulus figured by Bouvier,
1914, p. 288, fig. 6. (c.l. = 7*5 mm.), a. Anterior part of carapace, in dorsal aspect,
with left antennule and antennal peduncle, b. Right anterolateral spine with the two
spines immediately behind it, more magnified, c. Left antennule more magnified.
Jasus lalandei (H. M.-Edw.). Puerulus from " Discovery " Stn. 101, 15/X/26
[c.l. — 9-5 mm.), d. Anterior part of carapace in dorsal aspect, with left antennule
and antennal peduncle, e. Left antennule more magnified.

ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS 19

2 mm

- X ^C[

Fig, i.

20 ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS

2. Supraorbital spines not denticulate on upper margin ; plate formed by

these spines and the rostrum narrower and less advanced, most of

eye-stalks visible (Fig. la). Four pairs of spinules on thoracic

sternum. No median carina on carapace or abdomen. Carapace

spiny, three spines at each anterolateral angle and a longitudinal

row of spinules replacing each lateral carina. Second segment of

antennular peduncle shorter than third (Fig. ic). [Antenna not

quite i'5 times the body length and tapering distally. First pair of

pereiopods more robust than the others.] . . Palinurus Fabr.

b. Exopodite of maxilliped 3 short, without flagellum. No posterolateral spines on

thoracic sternum. Rostrum well developed (Fig. id). [Supraorbital spines

not denticulate on upper margin ; plate formed by these and the rostrum

narrow, moderately advanced, but most of eye-stalks visible. One spine at

anterolateral angle of carapace. No median, but two lateral carinae on

posterior two-thirds of carapace. Second segment of antennular peduncle

rather shorter than third (Fig. le). Antenna not quite 1*5 times the body

length and tapering distally. First pair of pereiopods more robust than the

others.] .......... Jasus Parker.

II. Antennular flagella at least as long as the peduncle, outer not much more robust than the
inner (Figs. 2a, $a, c). Anterior margin of antennular 1 tergum as wide as the space
between the tips of the supraorbital spines. Rostrum absent. The " longicornis "
group. [Usually a large pair of posterolateral spines on last somite of thoracic sternum
and three carinae on the carapace. One (occasionally two) spine (s) at anterolateral
angle of carapace. Second segment of antennular peduncle equal to third. First
pair of pereiopods not more robust than the others (Fig. 7b, c).] . Panulirus White.

a. THE PUERULUS STAGE OR POST-LARVA OF PANULIRUS sp.

(Figs. 2« ; 3 ; 7<0

Material. Taken in trawl by M.F.V. " Menika " on Leopold Reef ground,
Malindi, Kenya Colony, in 10 to 20 fms. — one specimen, presented by H. Copley,
Esq., Fish Warden, Nairobi. March, 1952.

Measurements : Length of body 21-5 mm. ; Length of carapace 8 mm. ; Length
of antenna (complete) 45 mm.

Description. The specimen had been preserved in formalin and the body,
though transparent in life, was apparently tinged with reddish brown. When
received it was complete except that pereiopods II and V on the right side were
missing. The most striking feature of the post-larva is the enormously long an-
tennae, measuring just over twice the body length. The flagellum ends in a flattened
oar-like expansion which is all the more conspicuous because the last 7 or 8 segments
of the spatula are brown in colour (Fig. 30 ; brown portion stippled). A narrow

1 See footnote on p. 18.

Fig. 2. a. Puerulus of Panulirus sp. from Kenya, in dorsal aspect, most of the antennal
flagella omitted, b. Distal portion of antenna of the last stage Phyllosoma of P.
homarus (L.) from the Marquesas, c and c'. Distal segments of incomplete antennal
flagella of two Pueruli of P. homarus (L.) from the Marquesas, d. Terminal segments
of antennal flagellum of the Puerulus of P. argus (Latr.) from Bermuda.

ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS 21

WJi^kM

3 mm.

ZOOL. II, 2.

Fig. 2.

§§

22 ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS

band of the brown colour, a trifle paler, is present about one-third of the way
along the flagellum (i.e., at distal end of proximal third).

Fig. 2a represents the post-larva in dorsal aspect, with the greater part of the
antennal flagella omitted. The carapace measures 8 mm. from the tip of the supra-
orbital spine to the posterior margin ; it bears only two pairs of spinules in addition
to that at each anterolateral angle. The anterior pair is situated immediately
behind the eyes ; the posterior pair is some distance behind, one spinule at the
anterior end of each lateral carina. If the specimen is examined carefully under a
binocular microscope, three pairs of incipient spinules can be detected situated
between the anterior and posterior pairs of spinules, i.e., two median (gastric) pairs
a short distance in front of the median carina, and another pair posterior to the
carinae on the supraorbital spines. The three longitudinal carinae on the posterior
two-thirds of the carapace are not very clearly seen in dorsal aspect, but in profile
they stand out clearly : Each lateral carina seems to be continued forwards to
the anterior pair of spinules ; the incipient spinules appear as thickenings of the
chitin, so that the median carina seems to bifurcate and to be continued forwards
to the base of each rostral horn. A pair of incipient spinules is present on the
antennular tergum, near the anterior margin. A thin carina, not visible in dorsal
aspect, runs parallel to the lateral edge of the carapace. The cervical furrow is
not very apparent at this stage.

The thoracic sternum is very similar to that of the Puerulus of Panulirus regius
B. Capello figured by Bouvier (1917, pi. ix, fig. 9) ; in addition to the pair of large,
backwardly-directed spines on the last somite, near the bases of pereiopods V,
there are blunt lobules, which diminish in size anteriorly, at the postero-lateral
angles of each of the three preceding somites. Bouvier's fig. 9 does not show these,
but that on the penultimate somite is visible in lateral aspect in his fig. 10. The
thoracic ganglia are clearly visible through the thin integument.

The abdomen and tail fan present no unusual features. No groove is present on
any of the terga parallel to the posterior margin, but the anterior margin of the
succeeding somite is visible through the thin chitin and is represented by a broken
line in Fig. 2a. The pleura of somites 2 to 6 each possesses a sharp, backwardly
directed spine which is not visible in dorsal aspect.

The antennule is short, extending only a little way beyond the antennal peduncle
(Fig. 2a) ; the external flagellum is the more robust, a trifle longer than the internal
and beset with special sensory setae on all but the proximal 6 or 7 segments. Each
flagellum has 18 or 19 segments. The details of the peduncle are shown in ventro-
lateral aspect in Fig. 3b and c, of the basal segment in dorsal aspect in Fig. 3d.
All the setae are very finely plumose, as are those forming what Gilchrist (1920,
p. 198, fig. 13) calls the " antennular screen." In the Puerulus of Jasus lalandei
(H. M. Edw.) Gilchrist shows these setae as heavily plumose. The basal segment

Fig. 3. Puerulus of Panulirus sp. from Kenya, a. Terminal portion of antennal flagel-
lum, to show the spatulate apex. b. Antennal and antennular peduncles, in ventro-
lateral aspect, c. The latter further enlarged, d. Basal segment of antennular
peduncle, in dorsal aspect, e. Pleopod on second abdominal somite, with most
of the long setae omitted, and apex of appendix interna further enlarged.

ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS 23

1mm.

1 mm.

Fig. 3.

24 ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS

of the peduncle is equal in length to the sum of the other two segments ; the stato-
cyst is conspicuous but without any statolith.

The details of the antennal peduncle are shown in Fig. 2a in dorsal and in Fig. 36
in ventrolateral aspects. The spatulate apex of the long flagellum is represented
in Fig. 3a, the shaded terminal segments being, in the preserved specimen, brown
in colour.

The eyes are large and conspicuous, extending as far as the anterolateral spinose
angles of the carapace.

The external maxillipeds are widely separated basally ; when extended they
reach the anterior margin of the antennular somite ; a minute trace of an exopodite
is present. The mouthparts were not dissected and examined in detail as I did
not wish to damage the specimen. But they appear to be very similar to those
of a Puerulus from the Marquesas represented in Fig. 6a-e except that the exopodite
on maxilliped 2 is rather shorter, being about three-fourths of the merus. The
mandibular palp is clearly seen without dissection.

Pereiopod I is appreciably shorter and somewhat more robust than the succeeding
limb (Fig. yc) ; the propodus is not quite three times as long as wide and the ratio
of dactylus to propodus is 1 : 1-25 ; the merus is rather longer than carpus and
propodus. Pereiopods II, III and IV are very similar in size and form, the propodus
in each being almost exactly five times as long as wide ; the ratio of dactylus to
propodus is 1 : 175-170 ; the merus equals propodus and carpus. The shrivelled
remnant of the exopodite at the base of pereiopods I — IV is very small and is best
seen when the limb (ischium to the tip) is detached, leaving the rudimentary exo-
podite exposed on the stump. Pereiopod V is the shortest and shows no trace of
the exopodite ; the propodus is 5 times as long as wide.

Pleopod 1, that attached to the second abdominal somite, is represented in Fig.
30 ; most of the long setae on the exopodite and endopodite are omitted ; the
appendix interna is long, and bears a subterminal plumose seta and a number of
terminal coupling hooks (see enlarged view of apex).

Remarks. Unfortunately the post-larvae of the Palinuridae are seldom complete
as to antennae, and this is especially true of those with long flagella. Several
Puerulus stages, however, have been figured with the antennal flagellum tapering :
e.g., that of Panulirus argus (Gurney, 1942, p. 234, fig. 93 ; Schmitt, 1935, p. 173,
fig. 36 of the young crawfish), and I have examined the specimen figured by Gurney
and find that the apex of the flagellum is as represented in Fig. 2d ; that of P. regius
B. Capello (Schmitt, 1926, p. 43, fig. 67) ; of P. japonicus (Nakazawa, 1917, plate,
fig s - 3> 4> an d 5 of the post-Peurulus). In these the antenna does not seem to
exceed, if it attains, 1-5 times the body length.

Dr. Holthuis of the Leiden Museum kindly informed me by letter that a Puerulus
with a long spatulate antennal flagellum had recently been described by Kubo in
a paper written almost entirely in Japanese, but with a short summary in English
(first paragraphs, p. 91 — Kubo, 1950). As no copy of this paper was available in
London, Dr. Holthuis sent me a photostat copy. (Later I received a reprint from
the author — see p. 28.) Kubo attributes this Puerulus to Panulirus " versicola "
(Latr.) — presumably a misprint for versicolor — and finishes his English abstract

ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS 25

as follows : " Some descriptions in regard to the puerulus of P. versicola have been
given by Ortmann (1894) [error for 1891], Caiman (1909), and de Man (1916). But
no one has given description on the signal feature of the antennae."

Kubo's Puerulus is certainly very similar to that from Kenya ; his figures are
small, and in the photostat copy some of the finer details are lost. The flagellum
is just over twice the body length and ends in a spatulate expansion which is
" stained with rather deep brown colour (in alcoholic specimens)," and I think his
figure shows the brown band about one-third of the way along the flagellum. But
I do not think that his Puerulus belongs to the same species as that from East
Africa described above. As a rule his specimens have two pairs of spines, one in
front of the other, at the anterior end of the lateral carinae and those behind the
rostral horns are well formed. The general impression one gets from Kubo's figures
is that the antennal and antennular peduncles are more robust than those of the
Kenya specimen — more like those of specimens from the Marquesas described
below (see p. 29 and Fig. 5a, b.). Nor do I think that Kubo is right in attributing
all his post-larvae to Panulirus versicolor (Latr.) ; at any rate those figured do
not agree with the specimens from Christmas Island which Dr. Caiman (1909, p. 444)
referred to that species (see p. 26).

I have found an earlier reference to a Puerulus with a spatulate apex to the
antennal flagellum in a paper by W. von Bonde (1930, p. 25). In his specimen
the antenna is also about twice the body length. The following extract may be
quoted from von Bonde's paper : " Note. — Only one of my Puerulus specimens
of Panulirus has the antennae intact. Each is developed at its distal extremity
into a large flattened oar-like structure. Such an antenna has been figured by
Richters (12) [error for (13)] in a large Phyllosoma (Ph. longicorne Guer., length
38 mm.)." Richters (1873, pi. xxxi, fig. 3) shows a last stage Phyllosoma with a
long antenna, the clavate apex of which is rather wider than that represented in
Fig. 2b and has no terminal spinose projection. The fragment represented in
Fig. 26 was found in the jar along with a number of post-larvae and young from
the Marquesas (see p. 29).

Thus several species of the genus Panulirus possess these spatulate antennae in
the post-larval stage. Leach's type specimens of Phyllosoma clavicorne were col-
lected in. West African waters ; von Bonde's post-larvae, which he referred pro-
visionally to P. burgeri de Haan (= homarus) y were from the East Coast of South
Africa ; others are now known from Japan (Kubo, 1950), Kenya and the Marquesas.
As von Bonde did not figure his specimens, I do not know from his description
whether the antennules are like those of the Marquesas material (Fig. 56), or more
slender as in the Kenya specimen (Fig. 3c). But, as they are stated to have two
pairs of branchial spines on the carapace they agree in this with the Marquesas
Pueruli and it is probable that they are referable to P. homarus (L.). In the absence
of post-Puerulus stages it is not possible to refer the Kenya specimen to the adult
species, since it is not known whether abdominal grooves are present or absent.
What can be said with certainty is that it belongs to a species with longer and more
slender antennules than either homarus or dasypus (if the latter is regarded as a
distinct species — see p. 35).

26 ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS

p. the POST-LARVA OF Panulirus versicolor (Latr.)

(Fig. 4)

Material. Christmas Island, Indian Ocean. Dr. C. W. Andrews coll. partly
on the reef and partly in crevices in the piles of the pier at Flying Fish Cove,
reg. no. 1909.5. 19.248-252. Five post-larvae.

(Also several later young crawfish stages, reg. no. 1909.5. 19, 243-247).

Descriptive note. These specimens have been identified by Caiman (1909,
p. 444), and his determination is fairly certain since he had older stages, up to
74 mm. in length of body, collected at the same time and place. In none of his
specimens is the antennal flagellum complete, but in one it exceeds 2-5 times the
body length and yet shows no trace of expansion distally — thus it must be consider-
ably longer than that of either the Kenya specimen described above, or of Kubo's
post-larvae referred to "P. versicola " (Kubo, 1950, p. 94, fig. 3). De Man (1916,
p. 61) says of his specimens " external antennae 3-times as long as the body " ;
this I at first regarded as a misprint for " 2-times " (as Kubo may also have done),
but Caiman's material suggests that de Man's statement may be correct. If de
Man's specimens were complete the apex of the flagellum was not spatulate or he
would certainly have said so. I have since examined de Man's specimens in the
Zoological Museum, Amsterdam. They certainly agree with our Christmas Island
material but in none are the antennae now intact.

In all the post-larvae there are only three pairs of spinules on the carapace — one
at the anterior end of the lateral carinae (not two, as in the majority of Kubo's
specimens), one behind the orbits and one behind the supraorbital spines ; in
addition two pairs of incipient spinules are present on the cardiac region. They
are thus similar to the specimen from Kenya except that the pair behind the rostral
horns is distinct. The antennal peduncle is long and slender, rather like but not
identical with that of the Kenya specimen (cf. Figs. 4a, b and 2a, 3c, d). The basal
segment differs in that the portion containing the statocyst is shorter while the
distal part is narrow and not expanded apically. A large round statolith is visible
within the sac (as in all the post -larvae examined other than the one from Kenya,
in which it may not have had time to form after a recent moult). Unfortunately,
in all the specimens the antennal peduncle is partially collapsed on the inner surface,
but it also is relatively long and slender. The pereiopods have also collapsed in
most of the specimens, but in one they are normal and I and III are represented
in Fig. 4c, at the same magnification as those of the other Pueruli. They are slightly
longer than those of the Kenya specimen (Fig. jc), or of the Marquesas material
described below (Fig. yb).

NOTE ON THE POST-PUERULUS AND LATER STAGES

As stated by Caiman (1909, p. 444), the post-Pueruli are brown in colour, with a
conspicuous W-shaped marking of white on the carapace and a transverse narrow
white band on each abdominal somite, that on somite 1 on the anterior half, those
on somites 2-6 parallel to the posterior margins. No abdominal grooves are present,

ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS

2 mm.

b k

Fig. 4. Puerulus of Panulirus versicolor (Latr.) from Christmas Island, a. Anterior
portion of carapace and antennal and antennular peduncles, in dorsal aspect, b.
Antennular peduncle in dorsal aspect, c. Pereiopod I (upper) and pereiopod III.

28 ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS

as in the adult, but on each somite a number of transverse rows of minute pits,
from each of which a short seta emerges, are present. A rather long forwardly
directed seta arises from the base of several of the median spines on the carapace.
After over 40 years in alcohol the colour pattern of these post-Pueruli is still
clear.

In the two older stages from Christmas Island (c.l. 25 and 34 mm. respectively)
the colour pattern is much nearer that of the adult ; the legs are striped, the dark
purplish blotches on the carapace tend to obscure the short median bands of the
W. The posterior whitish band across each abdominal somite is now flanked
anteriorly and posteriorly by a purplish band. What is rather surprising, however,
is the presence of a broad but very shallow transverse groove, with slight median
interruption, on abdominal somite 2. The pits and setae are now confined to this
area, the groove being filled with them while a few continue across the median
break. On somite 3 there is a much wider median interruption, beset with sparse
setae, between the two setose lateral grooves. On somite 4 there are just two short
widely separated lateral patches of setae. Any setae on 5 and 6 are isolated and
few in number. These broad shallow grooves on somites 2 and 3, or traces of
them, are present in rather larger stages from other localities in the B.M.
collection.

Remarks. As already mentioned on p. 25, I do not think that all Kubo's speci-
mens belong to P. versicolor (Latr.). The antennal flagellum in the Puerulus of
that species seems to be three times the body length and without a spatulate apex.
Since I commenced this paper I have received a reprint from Dr. Kubo and I find
that his fig. 4 o shows the W pattern rather faintly on the carapace of a young
stage (this was not apparent in the photostat copy). I now think that part of
Kubo's material only belongs to P. versicolor — namely the young with the charac-
teristic pattern and perhaps the Pueruli which he lists with but one spine at the
anterior end of each lateral carina (Kubo, 1950, p. 93, table 1). The other Pueruli
which he figures, with two spines at anterior end of the lateral carina and the stouter
antennal and antennular peduncles (Kubo, 1950, figs. 2 and 3) may belong to P.
homarus (L.) like my specimens from the Marquesas. I do not know how many
post-Puerulus and young stages Kubo had. But I assume that in his material no
abdominal grooves were present, since that is the case in P. versicolor. In the
post-Puerulus of P. homarus the abdominal grooves are present and show traces of
the crenulation as well.

If Kubo's Puerulus with the spatulate antennae really has no abdominal grooves
in the next stage, it might be referable to either P. ornatus (Fabr.) or P. polyphagus
(Herbst). The adult of the latter species has a long multi-segmented flagellum on
the exopodite of the second maxilliped as represented in Fig. 8/, and one would
expect a well formed flagellum in the Puerulus also — as in that of P. inflatus
(Bouvier) represented in Fig. ya. Kubo's Fig. 4 K seems to have no flagellum, or
a vestigial one as in fig. 6d perhaps. So that P. polyphagus would appear to be
excluded. On the other hand, P. ornatus does not appear to have been recorded
from Japanese waters, although it goes as far north as Formosa (Holthuis, 1946,
p. 141.).

ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS 29

c. the post-larva of Panulirus homarus (L.)
(Figs. 2b-d ; $a, b ; 6 ; yb)

Material, (a) Marquesas. " St. George " Expedition. W. 38.27.12.24. One
Puerulus.

(b) Marquesas ; sand of beach, Hana Hevane. C. Crossland coll. 1925-26.
Twenty four Pueruli, nine post-Pueruli and young crawfish stages up to 33 mm.
in body length.

Descriptive notes. The 25 Puerulus stages vary in body length from 22 to
24 mm. As each specimen has a narrow band of brown pigment some distance
along the antennal flagellum, I thought at first that they must be identical with
the specimen from Kenya described above. The antennal flagellum in some speci-
mens shows a hint of expansion of the segments at the broken distal end (Fig.
2c and c') and the flagellum in these is nearly twice the body length. My suspicion
that the apex of the flagellum must be spatulate was confirmed in a rather un-
expected way. Lying in the jar amongst detached limbs was the distal end of the
antennal flagellum of the last stage Phyllosoma represented in Fig. 26. The larval
flagellum is unsegmented, but within this can be seen the fully formed flagellum
of the Puerulus with the segmentation faintly visible through the chitin. It is
obvious that the apex of this flagellum would be spatulate. Presumably this
portion of the larval antenna broke off just before or during the critical moult from
larva to post-larva. The apex of the larval antenna is very similar to that of the
last stage Phyllosoma clavicorne Leach, the types of which are in the British Museum
collection. The antennal flagellum of these Puerulus stages, therefore, seems to
agree with that of the specimen from Kenya.

But on closer examination they were found to differ from the latter in a number
of characters. The resemblances are many, chiefly in characters common to nearly
all the Panulirus post-larvae that have been described such as the presence of the
posterior pair of spines on the thoracic sternum. The chief differences are as
follows : (i) In all specimens, even the smallest, four pairs of spinules are present
on the carapace in addition to the spinose anterolateral angles, namely, a post-
ocular pair, a postrostral pair, and two pairs one behind the other at the anterior
ends of the lateral carinae. Two pairs of incipient spinules are present on the
gastric region as in the Kenya specimen, (ii) The anterinules are more robust ;
the flagella have some 25 and 23 segments respectively ; the segments of the
peduncle are in the ratio of 2 : 1 : 1, as seems usual at this stage, but each segment
is broader relatively to its length (cf. Figs. 56 and 3c, d). The first segment is
2 -5 times as long as wide, the portion containing the statocyst is nearly as long as
the distal portion and a conspicuous statolith is always present. In the specimen
from Kenya the basal segment is 4 times as long as wide and no statolith was visible,
(iii) The antennal peduncle is shorter and more robust, (iv) The pereiopods also
seem to be somewhat stouter, especially as regards the merus (cf. Fig. yb and c).
The mouthparts are represented in Fig. 6a-e.

ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS

2 mm

1mm.

2 mm.

Fig. 5. Puerulus of Panulirus homarus (L.) from the Marquesas, a. Anterior part
of carapace with left antennal and antennular peduncles, in dorsal aspect, b. Anten-
nal peduncle. Puerulus of Panulirus inflatus (Bouvier) from Coiba. c. Anterior
part of carapace, antennal and antennular peduncles, in dorsal aspect, d. Antennal
peduncle.

ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS 31

Remarks. Because several later stages are represented it seems possible to
refer these young specimens to the adult species. These young crawfish possess an
abdominal groove on each somite, and, moreover, crenulations are present in the
older specimens in the lateral portions of each groove. The furrows are slightly
interrupted in the centre on the second and subsequent somites. The largest
young crawfish measures only 33 mm. from tip of supraorbital spine to tip of telson,
but it has a tiny anterodorsal spine on the merus, near the distal articulation, of
pereiopods II and III. The soft part of the telson is almost as wide as the hard
basal portion ; pereiopod III is longer than pereiopod II by half the length of
the dactylus ; the antennular peduncle is now very nearly equal to the antennal
peduncle, and the basal segment of the former reaches almost to the distal end of
the penultimate segment of the latter. Only one pair of slender spines are present
on the antennular tergum at this stage. Many spines are present on the carapace.
Holthuis (1946, pp. 110-112) lists the following species as widely distributed in the
Indo- Pacific region : Panulirus dasypus (Latr.), P. homarus (L.), P. japonicus
(von Siebold), P. ornatus (Fabr.), P. penicillatus (Oliv.), P. polyphagus (Herbst)
and P. versicolor (Latr.). Of these the species without abdominal grooves or furrows
may be excluded, leaving four — dasypus, homarus, japonicus and penicillatus.
The post-larva of P. japonicus has been described by Nakazawa (1917, figs. 3, 4,
and 5) ; he figures the antennae of the Puerulus as not exceeding 1-5 times the
body length and tapering distally. According to Bouvier (1913, p. 87, key) there
are no posterolateral spines on the thoracic sternum in "Puerulus pellucidus,"
which is referred to P. japonicus. In any case the antennule of the adult japonicus
(de Haan, 1833-49, pi- 4 I /4 2 ) has an unusually long slender basal segment to the
peduncle, equal in length to the sum of the three segments of the antennal peduncle,
and one would expect the specimens from the Marquesas to belong to a species
with a relatively short robust antennule. P. penicillatus can also be excluded
because (i) the margins of the abdominal grooves are never crenulate, and (ii) there
is in the adult a short but distinct exopodite on maxilliped 3 so that in the post-
larva one would expect the exopodite to be larger than it is in these specimens.
As the grooves on the abdominal somites are already distinctly crenulate laterally
and slightly interrupted in the median line I have no hesitation in referring these
young stages to either P. homarus or P. dasypus. Most authors are agreed that
these two species are very closely related and, in the Appendix, I give reasons for
believing that they belong to one rather variable species to which the older name
P. homarus (L.) should be given.

A short description, without figures, of the Puerulus and post-Puerulus stages
of Panulirus dasypus from the Red Sea was given by Bouvier (1913, p. 84 ; key to
the known Pueruli of the Palinuridae, p. 8y). These specimens were found " dans
les crevasses du rivage, a Djiboutil " (Bouvier, 1913, p. 88, and 1914, p. 191). These
young stages appear to be inhabitants of the littoral zone. The presence of crenula-
tions and slight interruptions on the abdominal grooves at this stages does not
necessarily mean that the grooves and sculpturing of the adults will remain of
the B-form (or dasypus form) — see p. 38.

32 ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS

Fig. 6. Puerulus of Panulirus homarus (L.) from the Marquesas, a. Mandibles, left
maxillule and lower lip. b. Maxilla, c. First maxilliped. d. Second maxilliped.
e. Third maxilliped (distal part of epipodite omitted).

ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS 33

d. THE POST-LARVA OF Panulirus inflatus (Bouvier)

(Figs. 5c, d ; 7a)

Material. Coiba Island. " St. George " Expedition 1924-25. One specimen
measuring 22 mm. in length.

Description. This specimen has the carapace more swollen and distorted
than usual. It agrees with most of the Pueruli that I have examined in having a
pair of posterolateral spines on the thoracic sternum near the bases of pereiopods
V. It differs from all the others in having two spines, one immediately behind the
other, on each anterolateral angle of the carapace (Fig. 5c) and a backwardly
directed spine at the posterior end of the median carina, i.e., on the posterior margin
of the carapace. In addition to the spinules shown in Fig. 5c, namely, at the antero-
lateral angles, behind the orbits and supraorbital spines respectively, and two
very minute pairs on the gastric region, there is only one at the anterior end of
each lateral carina.

The antennae are unfortunately incomplete, so that the nature of the apex is
unknown ; the longer exceeds 1-5 times the body length and shows no trace of
any expansion — which does not necessarily mean that the apex is tapering. The
details of the antennular and antennal peduncles are shown in Fig. 5c and d ; the
former is more robust than that of the Puerulus from Kenya, or of the specimens
referred to P. versicolor (Latr.). They recall those of the post- larvae from the
Marquesas described above ; the region of the statocyst is rather opaque but the
round statolith is visible. The mouthparts have not been examined in detail, but
there is just a vestige of the exopodite on maxilliped 3 (as in Fig. 6e). The exo-
podite of maxilliped 2 has a long flagellum, and in this respect differs from that of
the Marquesas material (cf. Fig. ya and 6d).

Remarks. According to Holthuis (1946, pp. no-in) there are but two species
from the Pacific Coast of Central America, namely, P. inflatus (Bouvier) and P.
interruptus (Randall). The latter species possesses in the adult the longest exo-
podite to maxilliped 3 of all Panulirus species and therefore its post-larva should
have an even longer exopodite than that of, e.g., P. argus (Latr.) figured by Gurney
(1942, p. 234, fig. 93E). Since the adult of P. inflatus lacks the exopodite on
maxilliped 3 and has a large multi-articulate flagellum on the exopodite of maxilliped
2, I have no hesitation in referring this Puerulus to that species.

Amongst the registered post-larvae in the B.M. collection there is another speci-
men from Esmeraldas, Ecuador (reg. no. 1925. 12. 8. 4), which also belongs to P.
inflatus. The carapace is normal, and again there is a double spine at each antero-
lateral angle, but the posterior spine, that at the distal end of the median carina,
is lacking. The antennal flagella are incomplete, and there is a rather broad band
of brown pigment, a considerable distance from the peduncle, on the left side only.
In both specimens there are two rows of 3-4 minute spinules on the proximal half
of the soft part of the telson ; no spinules are present in the post-larva from Kenya
I (Fig. 2a) ; in the Pueruli of P. homarus there may be two pairs of minute spinules

ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS

in some of the specimens ; a few spinules are also present in some of the post-
larvae of P. versicolor.

The holotype of " Puerulus inermis " Pocock, said to be the post-larva of Panulirus
guttatus (Latr.) (Holthuis, 1946, p. no), has the antennular peduncle of the shorter

Fig. 7. Puerulus of Panulirus inflatus (Bouvier) from Coiba. a. Right maxilliped 2.
Puerulus oi Panulirus homarus (L.) from the Marquesas, b. Pereiopods I, II and
III (from left to right). Puerulus of Panulirus sp. from Kenya, c. Pereiopods I
(above), II and III.

broader form found in P. homarus and P. inflatus (Fig. $a-d). The exopodite of
maxilliped 2 has a shorter flagellum than that of the Puerulus of P. inflatus (Fig.
ya) and does not reach beyond the distal third of the carpus of the endopodite. The
exopodite of maxilliped 3 is vestigial, as in Fig. 6e. The antennal flagellum is

ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS 35

incomplete. Pocock's specimen was dredged in about 10 fms. in Water Bay,
Fernando Noronha (Pocock, 1890, p. 516).

APPENDIX

Note on Variation of the Adults of Panulirus homarus (L.)

(Figs. 8 ; 9)

Two Indo-Pacific species of Panulirus, namely P. homarus (L.) — better known
as P. burgeri (de Haan) — and P. dasypus (Latreille), are apparently very closely
related. The differences between them are indicated in the following extract
from Gravel's key (191 1, p. 50) :

" Exopodite des deuxieme maxillipedes avec fouet atrophie. Sillons des tergites
abdominaux toujours denteles et nettement ininterrompus

4. P. burgeri, de Haan.
" Exopodite des deuxieme maxillipedes sans fouet. Sillons des tergites abdominaux
denteles et toujours au moins legerement interrompus sur la ligne mediane

5. P. dasypus, Latreille."

Recently Holthuis (1946, pp. 128-136) has attempted to find clear-cut characters
by which the two species may be separated ; he lists only some minor differences,
and sums up by saying that more abundant material may perhaps show that the
two are identical. In recent years I have had considerable correspondence with
Dr. Leakey, of the Coryndon Museum, Nairobi, relating to the East African species,
and from time to time he has sent me a few specimens for determination. Some
of these seemed to agree with P. homarus, others with P. dasypus as far as the
abdominal grooves were concerned, but, as Dr. Leakey pointed out, not always as
regards the second maxilliped. For example, two specimens with the deep crenula-
tion and uninterrupted furrows typical of P. homarus differed as to the exopodite
of maxilliped 2 ; in one a small one-segmented flagellum was present as in Fig. Sd,
whereas in the other the flagellum was absent as in Fig. 8c. The former therefore
agreed with Gruvel's definition of P. burgeri (avec fouet atrophie), while the latter
agreed with that of P. dasypus (sans fouet). Two other specimens with the crenula-
tion much less marked, especially mid dorsally, differed as follows : one was of a
reddish colour, with uninterrupted grooves and no flagellum on the exopodite of
maxilliped 2 (as in Fig. 8c) ; the other was bluish green, with a short but distinct
median interruption of the groove on somites 2, 3 and 4 respectively, whereas the
exopodite of maxilliped 2 has a large flagellum. (In 1948 I did not observe that
the flagellum differed on right and left sides : that on the right is represented in
Fig. 8a ; the flagellum can scarcely be described as " atrophie," being half the
length of the basal segment of the exopodite and rather irregularly segmented ;
that on the left is as represented in Fig. 8d and " atrophie.")

I have now re-examined the spirit specimens in the British Museum collection
that have been referred to P. burgeri (de Haan) and P. dasypus (Latreille) by Dr.
Caiman, who revised the material of Panulirus at one time (unpublished), and by

36 ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS

3mm.

Fig. 8. Portions of the second maxilliped, in anterior aspect, to show the exopodite
in various species of Panulirus. a-f. Adults, g and h very young specimens.

Panulirus homarus (L.). a. From specimen no. 3 in Table, with distal part of
exopodite from the side normally facing the endopodite. b. Distal part of exopodite
of right maxilliped of specimen no. 9. c. Left maxilliped of specimen no. 4. d.
From specimen no. 6. e. From specimen no. 1. g. Right maxilliped of largest
young stage from the Marquesas (c.l. = 13 mm.).

Panulirus polyphagus (Herbst). /. Left maxilliped of a $ [c.l. = 66 mm.) with
the flagellum, in inner aspect, more enlarged.

Panulirus versicolor (Latr.). h. Right maxilliped of a young specimen (c.l. = 25
mm.) from Christmas Island.

ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS

myself. Unfortunately there are only fifteen specimens, three of which were
referred by Dr. Caiman to dasypus and eight to biirgeri, the remaining four being
those already mentioned from East Africa. Table I shows briefly the variation in
(a) the abdominal grooves and the sculpturing in their vicinity, and (b) the exopodite

Fjg. 8— e-h.

ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS

of the second maxilliped, exhibited by the twelve specimens which obviously belong
to the biirgeri-dasypus complex. At first I thought that the six specimens with the

TABLE I

Panulirus homarus (= biirgeri) — specimens in B.M. Collection. 1

Abdominal

No.

Sex

c.l.
in

Reg. no.

grooves

Exopodite

of second maxilliped

mm.

Locality

and
sculpturing

Flagellum

Form of flagellum

• <J

105

1906.5.29.36
Socotra

. A ++ .

Absent

R.
L

Fig. 8e.

As in Fig. 8e.

• 6*

c. 90

1948.3. 12. 1

B +

}

Alike — as in

E. Africa

Fig. 8c.

. 6*

c. 88 .

1948.3. 12.2

B +

Present

Fig. 8a.

E. Africa

2, 3. 4

(dissimilar)

As in Fig. 8d.

. 6*

Kenya, 1950

. A +

Absent

R.
L.

}

Alike — as in
Fig. 8c.

• 6*

1925.8. 18.87

»»

{

Ditto.

Off Natal

. ?

Kenya, 1950

. A +

Present
(similar)

R.
L.

Fig. 8d.

As in Fig. 8d.

• ?

28/xi/o8

B +

Present

Fig. ga.

ovig.

No data

2, 3

(dissimilar)

Fig. ga'.

• ?

1925.8. 18.86

A +

As in Fig. 8e.

juv.

Off Natal

Absent

Like R, but
slightly
shorter.

. $

1910.3.29.36

Present

Fig. 8b.

juv.

Goram Id.

2, 3

(similar)

As in Fig. 8b.

juv.

1928. 12. 1. 36
S. Africa

Absent

R.
L.

}

Alike — as in
Fig. 8c.

• 6*

80.6

Present

Alike — as in

juv.

Amboina

(similar)

Fig. 8d.

• 6*

820a

B -

R. absent

As in Fig. 8c.

juv.

Ceylon

3> 4

L. present

As in Fig. 8d.

A ++• Crenulation and sculpturing very pronounced, of the " megasculpta " form.
A + . Intermediate between A and A + + ; some additional sculpturing and crenulation on
certain somites.

A. Crenulation well marked along entire anterior margin of each uninterrupted groove.

B. Crenulation moderately developed, most marked laterally, on anterior margin of groove.
B +. As for B, but with traces of pitting on somites 2 and 3, or 2-4.

B — . Crenulation faintly indicated laterally, fading out medially. Some specimens of the
B-form show a slight median interruption of the grooves on the somites indicated below B.

R and L = right and left second maxilliped respectively.

c. I. is measured from tip of supraorbital spine to posterior margin ; in specimens 2 and 3
the supraorbital spines are damaged.

1 Recently, in the Zoological Museums of Leiden and Amsterdam, I have examined the material
referred to P. homarus and P. dasypus, and find that it is just as variable as regards mouthparts and
abdominal grooves as the B,M. material,

ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS 39

A-form of grooves and sculpturing belonged to biirgeri, the other six with grooves
and sculpturing of the B-form to dasypits. In five of the former the flagellum of
maxilliped 2 is absent, whereas in five of the latter it is present on both sides or
(in one instance) on one side only. But the fact that in three out of the twelve
specimens the exopodites are dissimilar on right and left maxillipeds (nos. 3, 7 and
12), and that the three forms of exopodite are thus combined (flagellum present and
either long and 2- to 3-segmented, or short and 1- segmented, or absent), proves
that this distinction is untenable. I am forced to conclude that these twelve
specimens belong to a single variable species, P. biirgeri (de Haan), which according
to Holthuis must now be called P. homarus (L.). In all probability many, if not
all, of the references to P. dasypus in the literature refer to the less crenulate form
of P. homarus (this difference does not depend on size as far as one can see). Speci-
men no. 1 in the Table, from Socotra, is that referred to as P. dasypus in Pocock,
I 9°3» P- 2I 4 i the sculpturing on the abdomen is very pronounced and of the
" megasculpta " form and I have no hesitation in referring it to biirgeri. The exo-
podite of maxilliped 2 is as long as the merus and yet there is no trace of a suture
near the apex (Fig. 8e) ; Monod & Petit (1929, fig. 3E) figure the exopodite as
only half the length of the merus in their P. dasypus, so that the range of variation
here is considerable. As this is the largest specimen, the spines on the carapace
are rather worn, but they are not as few in number as in Gruvel's specimen of
dasypus (1911, pi. ii, fig. 5). The anterodorsal spine on the merus of pereiopods II
and III, with one exception, have almost entirely vanished, so I do not think that
the presence or absence of these spines is of specific importance.

Another difference listed by Holthuis (1946, p. 135), namely that pereiopods II
and III are about equal in homarus whereas III is appreciably the longer in dasypus,
is, I think, capable of another explanation. In small specimens of P. homarus,
irrespective of sex, the pereiopods are relatively short. In the juvenile from Goram
Island, for example, pereiopod I reaches the distal end of the second segment of the
antennal peduncle, II is just a trifle longer and Ilia trifle longer still, reaching the
distal end of the third segment of the peduncle ; IV reaches the middle of the
propodus of I. By the time the carapace has increased to 88 mm. in length (no. 3
in Table) all the legs are a little longer relatively to the antennal peduncle, but the
difference between II and III is not marked ; I exceeds the second segment of
the peduncle by its dactylus, while III exceeds I by rather more than its dactylus
(by 24 mm.) and IV is now a trifle longer than I. In the largest specimen (no. 1)
pereiopods I and IV are subequal and reach the distal end of segment 3 of the
peduncle much as before, but II exceeds I by the dactylus and one-fourth of the
propodus (33 mm.), while III, now clearly the longest, exceeds II by its dactylus
(30 mm.) and I by 63 mm. This specimen thus exhibits the commencement of a
feature peculiar to senile males at least. The difference gets more pronounced with
age, and this may well account for the long third pereiopod in Gruvel's specimen
of P. dasypus (1911, pi. ii, fig. 5). Gruvel does not mention the sex of the specimen
in the legend to the figure, which is one-fourth natural size, giving a c.l. of about
200 mm. Recently, when I had occasion to name some large dried specimens
for the Zanzibar Museum, I was much impressed by an extraordinary case of alio-

4 o

ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS

metric growth of pereiopods II and III in senile males of Panulirus ornatus (Fabr.).
Perhaps the females also exhibit the same phenomenon in old age. but the speci-
mens at my disposal happened to be males. Two that are now in the B.M. collec-
tion show the commencement of the differential rate of growth of pereiopods II
and III, and a much later stage, respectively. In the smaller the c.l. is 134 mm.

Fig. 9. Panulirus homarus (L.). a and a'. Right and left maxilliped 2 of specimen
no. 7 in Table.

Panulirus ornatus ? (Fabr.). b. Right maxilliped of $ from Shanghai (c.l. = 76
mm., reg. no. 62.95). c - Same of $ from Swatow (c.l. = 72 mm., reg. no. 84.2).

ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS 41

and greatest c.b. is 85 mm. ; pereiopods I and IV are of equal length and probably
scarcely reach beyond the distal end of the second segment of the antennal peduncle 1 ;
pereiopod II exceeds I by rather more than the dactylus (32 mm.) ; pereiopod III
exceeds I by the dactylus and some two-fifths of the propodus (55 mm.). The
larger male with c.l. 170 mm. and greatest c.b. 130 mm. must have been about
twice the live weight of the smaller one. Pereiopods I and IV are still equal and
relatively short ; pereiopod II exceeds I by nearly all its propodus and the dactylus
(about 115 mm.), while III exceeds I by some 186 mm.

Perhaps all species of the genus Panulirus will prove to undergo the same striking
increase in length of some of the pereiopods in old age.

A well-developed exopodite, with long multiarticulate flagellum, on the second
maxilliped is characteristic of some species, e.g., Panulirus polyphagus (Herbst) —
unless, on re-examination of many specimens, the flagellum proves to be variable.
This type of exopodite is presumably primitive, and there has been a tendency to
reduction and loss of the flagellum within the genus. P. homarus would appear to
be in process of losing the flagellum, which would explain the variation exhibited
by the few specimens available. There are some dried specimens in the collection,
but they are not suitable for a study of the mouthparts.

Finally, to return to the three specimens determined as P. dasypus by Dr. Caiman,
I am rather at a loss to know what to call them. They comprise two females from
Shanghai and Swatow, China (c.l. 76 and 72 mm. respectively), and a young female
from Palau Satang. In the juvenile the mouthparts are damaged ; the older
specimens agree in having the exopodite of the second maxilliped shorter than the
merus of the endopodite, and more slender than that of P. homarus (Fig. gb, c ;
c.f., Fig. 8c). The carapace is quite spiny and, in all three, abdominal grooves
are present on somites 2 and 3, absent or evanescent on the posterior somites.
There may be a median break or interruption, but the grooves are wider and shallower
than in any of the homarus specimens examined, and filled with a fine short pubes-
cence. When this is rubbed off one might overlook the grooves entirely and place
the specimens in the subdivision without abdominal grooves. There is no trace of
the crenulation which is always referred to in connection with P. dasypus and the
antennular peduncle is relatively longer and more slender than in any of the speci-
mens tabulated under P. homarus. These three specimens do not seem to agree
with those referred to dasypus by Gruvel (191 1, pi. ii, fig. 5), Monod & Petit (1929,
p. 277, fig. 3 a-e) and others, and which appear to be the B-form of the variable
P. homarus discussed above. Young specimens of P. versicolor have a broad shallow
groove, filled with pubescence and with a median interruption, on somites 2 and
3, and just a trace of the setae on each side of somite 4 (see p. 28). These three
specimens may belong to another species with abdominal grooves absent, as a
rule, in adults — such as P. ornatus (Fabr.). Incidentally, there are records of
P. ornatus from Swatow and Poeloe Satang, which is doubtless the same as Palau
Satang — see Holthuis, 1946, p. 141. P. ornatus attains a very large size and might
retain a juvenile character like these very shallow grooves for a considerable time.

1 The pereiopods are all directed forwards and close together, while the antennae are bent back, as
convenient for storing dried material, so that exact comparison is not possible.

42 ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS

These specimens are faded, but in the larger ones there are traces of the broad band
of colour present on each abdominal somite of large specimens of P. ornatus.

REFERENCES

Bonde, W. von. 1932. Post-brephalus development of some South African Macrura. Rep.

Fish. Mar. Biol. Surv. S. Africa, 8, Special Rept. no. 1, 1930 [1932] : 1-42, 14 pis.
Bouvier, E. L. 1913. Le Stade " Natant " ou " Puerulus " des Palinurides. Trans. 2nd

Intern. Congress Entomology, Oxford, 1912 [191 3], 2 : 78-89.

1914. Recherches sur le developpement post-embryonnaire de la Langouste commune

(Palinurus vulgaris). J. Mar. Biol. Ass. Plymouth, n.s. 10 (2) : 179-193, 6 text-figs.

191 7. Crustaces decapodes (Macroures marcheurs) provenant des campagnes des yachts

" Hirondelle " et " Princesse-Alice " (1885-1915). Res. camp. set. Monaco, Paris, 50 : 1-140,
11 plates.

— — 1925. Reports of the Results of Dredging ... by the " Blake." XLVIII : Les
Macroures Marcheurs. Mem. Mus. Comp. Zool. Harvard, 47 (5) : 401-472, 11 plates.

Calman, W. T. 1909. The genus Puerulus Ortmann, and the post-larval development of
the Spiny Lobsters (Palinuridae) . Ann. Mag. Nat. Hist. (8) 3 : 441-446.

Gilchrist, J. D. F. A post-puerulus stage of Jasus lalandii (Milne Edw.), Ortmann. /.
Linn. Soc. Lond. Zool. 34 : 189-201, pis. 15 and 16, 13 text-figs.

Gruvel, A. 191 1. Contribution a l'etude generale systematique et economique des Pali-
nuridae. Ann. Inst, ocean. Monaco, Paris, 3 (4) : 5-56, 6 pis.

Gurney, R. 1942. Larvae of Decapod Crustacea. Ray Society, vol. 129 (for the year 1941),
pp. viii -f- 306, 122 text-figs. London.

Haan, W. de. 1833-50. Crustacea. [In P. F. de Siebold's Fauna Japonica] : 1-244, P* s -
i-55 + A-Q.

Holthuis, L. B. 1946. The Decapoda Macrura of the " Snellius " Expedition : I. Tern-
minckia, 7 : 1-178, pis. i-xi.

Kubo, I. 1950. On two forms of Puerulus found in Japanese waters, with special reference
to that of Panulirus versicola (Latr.). Bull. Japan Soc. sci. Fish., 16 (3) : 91-98, 4 text-
figs., 2 tables.

Man, J. G. de. 1916. The Decapoda of the " Siboga " Expedition. Part III. Families
Eryonidae, Palinuridae, Scyllaridae and Nephropsidae. Siboga-Expeditie. Monographic,
xxxix a 2 , Leiden, 122 pp., 4 pis.

Monod, Th., & Petit, G. 1929. Contribution a l'etude de la Faune de Madagascar. Crus-
tacea. II. Palinuridae. Faune des Colonies Francaises, 3 (4) : 269-291, 9 text-figs.

Nakazawa, K. 191 7. On the metamorphosis of Panulirus japonicus. Dobuts Z. Tokyo,
29 : 259-267, 1 pi.

Ortmann, A. 1891. Die Decapoden-Krebse des Strassburger Museums . . . III. Die
Abtheilungen der Reptantia Boas : Homaridea, Loricata und Thalassinidea. Zool
Jahrb., Jena, Abth.f. Syst. 6 : 1-58, 1 pi.

Pocock, R. I. 1890. Crustacea. [In Notes on the Zoology of Fernando Noronha, by H. N.
Ridley.] /. Linn. Soc. Lond. Zool., 20 : pp. 506-526.

1903. The Decapoda of Sokotra. [In The Natural History of Sokotra and Abd-el-Kuri,

edited by Henry O. Forbes. Special Bull. Liverpool Mus. : 212-215, 1 text-fig.]

Richters, F. 1873. Die Phyllosomen. Ein Beitrag zur Entwicklingsgeschichte der Lori-

caten. Zs. wiss. Zool. Leipzig, 23 (4) : 623-646, pis. 31-34.
Schmitt, W. L. 1926. The Macruran, Anomuran, and Stomatopod Crustaceans collected

by the American Museum Congo Expedition, 1909-1915. Bull. Amer. Mus. Nat. Hist.,

53 (1) : 1-67, 75 text-figs., 9 pis.

I 935- Crustacea Macrura and Anomura of Porto Rico and the Virgin Islands. Sci.

Surv. P. Rico (New York Acad. Sci.), 15 (2) : 125-227, 80 text-figs.

PRESENTED

PRINTED IN GREAT BRITAIN BY
ADLARD AND SON, LIMITED,
BARTHOLOMEW PRESS, DORKING

1 NOV 1953

ON SIRPUS, A GENUS OF
PIGMY CANCROID CRABS

ISABELLA GORDON

BULLETIN OF

THE BRITISH MUSEUM (NATURAL HISTORY)

ZOOLOGY VoL 2 Na 3

LONDON : 1953

ON SIRPUS, A GENUS OF PIGMY
CANCROID CRABS

ISABELLA GORDON

Pp. 43-65 ; 13 Text-figures

BULLETIN OF

THE BRITISH MUSEUM (NATURAL HISTORY)

ZOOLOGY Vol. 2 No. 3

LONDON : 1953

THE BULLETIN OF THE BRITISH MUSEUM
(NATURAL HISTORY), instituted in 1949, is
issued in five series, corresponding to the Departments
of the Museum, and an Historical Series.

Parts appear at irregular intervals as they become
ready. Volumes will contain about three or four
hundred pages, and will not necessarily be completed
within one calendar year.

This paper is Vol. 2, No. 3 of the Zoological Series.

PRINTED BY ORDER OF THE TRUSTEES OF
THE BRITISH MUSEUM

Issued November 1953 Price Six Shillings

ON SIRPUS, A GENUS OF PIGMY
CANCROID CRABS

By ISABELLA GORDON, D.Sc, Ph.D.

SYNOPSIS

The paper contains an emended diagnosis of the genus Sirpus Gordon, a fuller description
of the type species 5. zariquieyi from Spain, and a description of a new species, S. monodi, from
West Africa. Both species are of very small size, though some of the specimens are sexually
mature. The affinities of the genus are discussed ; while it bears some resemblance to Trachy-
carcinus Faxon, it seems to be much nearer to Pirimela Leach and is referred provisionally to
the family Pirimelidae.

INTRODUCTION

In September, 1947, I received from Dr. Ricardo Zariquiey Alvarez of Barcelona
two tiny specimens of what appeared to be a species of Spider Crab from Cadaques
on the north-east coast of Spain. Although the specimens were minute it was
obvious that they were females, but I did not at that time observe that the larger
one has the genital openings perforate and filled with what appear to be fertilization
plugs. Later Dr. Zariquiey sent me a somewhat larger ovigerous female and two
males and Dr. H. Boschma of the Leiden Museum lent me a male specimen, all
from the same locality. For a time I was much puzzled by this pigmy species of
Brachyuran ; while at first sight it seemed to belong to the Oxyrhyncha or Spider
Crabs, it proved on closer examination to belong to the Cancroid assemblage, in
the wide sense employed by Rathbun, 1930. But its position in that assemblage
was not easy to determine, so I consulted Dr. H. Balss, formerly of the Munich
Museum, and Dr. Th. Monod of the Institut Francais d'Afrique Noire, Dakar. Shortly
afterwards Dr. Monod wrote to say that he was sending me some specimens from
West Africa that he had found among his undetermined material and which seemed
to belong to the same genus. They proved to represent another new species which
is described in this paper. Pending the arrival of Monod's specimens I wrote a
preliminary paper describing the Mediterranean species under the name of Sirpus
zariquieyi n.g. and sp., and discussed very briefly the affinities of the genus. Now
that a second species is available I am able to give an emended diagnosis of the
genus; more figures, a fuller description and details as to habitat of the type species,
and a detailed account of my reasons for referring the genus to the family Piri-
melidae are also included.

Genus Sirpus Gordon
Gordon, 1953 > P- 3°4-

Diagnosis (emended). Carapace hexagonal, convex, areolate, the lobes bearing
granular clusters or nodules. Front rather narrow, produced, trispinose, the spines
or teeth diverging from a common base and thus not in one plane. Orbits large,

ZOOL. 11, 3. 3

ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS

with forward aspect, shallow and imperfect posteroventrally ; two gaps in dorsal
margin, a wide one below and one at the inner angle. Eyestalks long, cylindrical,
with a bulbous base, retractile within the shallow orbits ; eyes deeply pigmented
and protected by the postocular or first anterolateral spine. Anterolateral and
posterolateral margins about equal, the former with four spines or teeth, the third
of which is the smallest. Antennal flagellum long and setose ; basal segment
(2 + 3) of peduncle but slightly enlarged, not quite filling hiatus at inner angle of
orbit but attaining the front, subcylindrical and armed at outer distal angle ; seg-
ments 4 and 5 subequal. Antennules folded rather obliquely. Anterior margin
of buccal cavity distinct. Sternum long and rather narrow. Abdomen rather
narrow in both sexes and terminal somite triangular, longer than wide ; all seven
somites free in female, three to five coalesced in male. No appreciable gap between
external maxillipeds ; a rounded lobule at antero-internal angle of merus and antero-
external angle rounded ; antero-internal angle of ischium not produced. Chelipeds
equal, carpus and palm of chela nodular and/or granulose. Walking legs rather
short, one longitudinal dorsal ridge on carpus, two on propodus ; dactyli not com-
pressed but longitudinally ridged, longer than propodus. Male openings coxal ;
first pleopod a long slender style, about four times as long as the second.

As I explained in the preliminary paper, I have used the less usual spelling of
scirpus in deference to the wishes of my botanical colleagues because Scirpus is
a large and well-known genus or plants. According to Lewis and Short, A Latin
Dictionary, Oxford, 1917, scirpus, sometimes sir pus, has two meanings: (i)
a rush or bulrush, and (ii) by transference " deriving the idea of intricacy from
plaited work of rushes, a riddle, enigma," and I use it in the latter sense. The
gender is masculine.

Two species are known at present, the type species Sirpus zariquieyi Gordon
from Cadaques, and Sirpus monodi n. sp. from Dakar. The affinities of the genus
appear to be with Pirimela and T r achy car cinus, perhaps nearer to the former —
see p. 59-63.

Sirpus zariquieyi Gordon
Gordon, 1953, p. 305.

Diagnosis as for the genus. In this species there is considerable sexual dimorphism,
the carapace in the female being distinctly wider and more ovoid than in the male
(cf. Figs, ia, 4 and 5).

Material. From Dr. Zariquiey : (a) Two small females collected from fisher-
men's nets while being cleaned ashore, but from 20 to 40 metres in the neighbour-
hood of Cadaques, Provincia de Gerona, Spain. Summer, 1947.

(b) One ovigerous female (the holotype) and two small males from Cystoseira
growth on the rusty iron plates of a sunken ship, in about 20 metres, off Cadaques,
viii.51.

From Leiden Museum : (c) One small male, from a fishing net and caught in the
vicinity of Cadaques, 4th-i6th August, 1949.

ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS

Dr. Zariquiey has very kindly presented the holotype and one male specimen to
the British Museum Collection.

Description of ovigerous female. This specimen, the largest at my disposal,
measures almost exactly 5 mm. from posterior margin of carapace to apex of rostrum
and 7- 1 mm. in greatest width (between the last pair of anterolateral spines).

mm,

Fig. 1. Sirpus zariquieyi Gordon, a. Holotype, an ovigerous female, in dorsal aspect,
pereiopods and right antenna omitted, b. Left chela and carpus of holotype. 1.
Supraorbital spine. 2. Infraorbital spine.

The form of the carapace is represented in dorsal aspect in Fig. ia. The front
consists of three subequal spines ; the median one, or rostrum, is directed obliquely
downwards, while the lateral ones are directed forwards and are slightly upturned
apically. The orbits are wide, so that the distance between the postocular spines
(first pair of anterolateral spines, Fig. ia) exceeds two-thirds of the maximum cara-

46 ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS

pace width. Each orbit is well formed anteriorly but is shallow and incomplete
ventro-posteriorly, so that the retracted eye is protected only by the postocular
spine. The supraorbital and infraorbital angles are each produced to form a short
triangular spine (Fig. ia, 1 and 2 respectively). The dorsal orbital margin slopes
obliquely backwards and outwards to meet the base of a wide triangular spine inter-
calated between it and the postocular spine. Behind the orbit the anterolateral
margin comprises three large spinose projections, the anterior of which is the largest.
The posterolateral margins are slightly concave and convergent ; the posterior
margin bears a high submarginal crest, and is only a trifle wider than the trispinose
front.

The dorsal surface of the carapace is distinctly areolated as represented in Fig.
ia, and the apex of each lobe is beset with some granules. The gastric region has
three lobes, two protogastric and one mesogastric — the latter followed by a cardiac
and a small intestinal lobe. On either side of the cardiac lobe is a dark reniform
depression. The branchial region bears a median lobe (slightly behind and external
to the mesogastric lobe), a triangular lobule near the base of the posterior antero-
lateral spine, and two spinules and a few granular rugosities near the posterolateral
margin (Fig. ia). A number of rather long plumose setae are present as indicated
in the figure, but others may have been rubbed off. The anterior portion of the
carapace is represented in profile in Fig. 2B.

The antennae are long and setose ; the ventral spine on the outer distal angle of
segment 2 + 3 of the peduncle is visible in dorsal aspect ; the fourth segment is a
trifle longer than the fifth (Fig. ia, see also Fig. 2A). The flagellum is a trifle more
robust than I have indicated in the figure and comprises some 14-16 segments,
some of which bear long, distally plumose setae.

The eyestalk is much swollen basally, cylindrical in the distal two- thirds, the
corneal area is terminal.

The chelipeds are equal ; the distal segments of the left one are represented in
Fig. ib. As the carapace is wider than in the male, the merus reaches the apex of
the third anterolateral spine and only a small portion is visible dorsally — cf . Fig. 4
where more of the merus is visible beyond the anterolateral margin. The spine on
the inner border of the carpus is much longer than in the male (which is of smaller
size) ; on the anterior margin a blunt lobule articulates with a similar one on the
chela ; the distal half of the outer margin is cristate or lobulate and beset with
granules ; on the dorsal surface are two inner, and two or three outer clusters of
granules. The fingers of the chela are rather longer than the dorsal margin of the
palm ; each bears a median crest or ridge ; the cutting edges are in close contact
distally, leaving a slight gap proximally. The dorsal margin of the palm bears a
large distal and a smaller median granular prominence, and there are two similar
prominences a short distance from the actual margin. Below these is a median
series of low granular clusters. The walking legs or pereiopods II-V are for the most
part bent inwards round the mass of ova and I did not attempt to straighten them
out, so have omitted them from the figure. The first pair are about as long as the
chelipeds. The right pereiopod V is represented in Fig. 3B ; the merus is about
two and a half times as long as wide, the carpus bears a low dorsal ridge near which

ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS

is a distal depression. The propodus, which is as long as the carpus, has two parallel
dorsal ridges ; the dactylus is half as long again as the propodus and bears longi-
tudinal ridges ; in none is it compressed.

I mm.

Fig. 2. Tr achy car cinus glaucus Ale. & And.
syntype (c.l. = i6-i, c.b. = 14-8 mm.)

a. Abdominal somites 3 to 7 of male

Sirpus zariquieyi Gordon, b. Left anterior portion of carapace of larger para type
from Lot a, in lateral aspect, to show divergence of the frontal spines, c. Abdominal
somites 3 to 7 of same specimen, d. Abdominal somites 3 to 7 of male para type.
j. Lateral frontal spine. 2. Median frontal spine. 3. Infraorbital spine. 4. Supra-
orbital spine. 5. First and second anterolateral spines.

zool. 11, 3.

3§

4 8

ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS

The antennules are enclosed in deeper, more developed sockets than in the male
represented in Fig. 3A and the crest on the anterior margin of the basal segment of
the peduncle is more pronounced. The separation between the base of the rostrum
and the epistone is also more pronounced in the female, and the median notch on
the anterior margin of the buccal cavity is less distinct. The longitudinal furrow
on the ischium of maxilliped 3 is also more pronounced than in the male.

Fig. 3. Sirpus zariquieyi Gordon, a. Right anterior portion of carapace of larger male
paratype from Lot b, in ventral aspect, to show details of orbit, epistome, buccal cavity,
etc. b. Right pereiopod V of holotype. 1. Supraorbital spine. 2. Infraorbital spine.

All seven segments of the abdomen are free ; the first segment is very short and
has a transverse granular ridge ; the next three are progressively longer while the
last three are of equal length, the terminal one being narrowly triangular with a
rounded apex. The abdomen is comparatively narrow, since at its widest it scarcely
exceeds the width of the posterior border of the carapace. Somites 3 to 7 are repre-
sented in Fig. 2C.

The ova are numerous, and measure approximately 0-25 to 0*30 mm. in diameter.

Remarks. The larger female from Lot a measures 4-85 by 5-52 mm. and is
very similar to the holotype. It appears to be sexually mature, for the genital
openings are distinct and seem to be filled with dark brown fertilization plugs. The
clusters of granules on the carapace are more pronounced, and from them arise

ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS

groups of distally plumose setae. The swollen basal part of the eyestalk is also
closely beset with small granules and is separated from the distal portion by a distinct
depression. The smaller female from Lot a measures 4-00 by 4-81 mm. and is

Fig. 4. Sirpus zariquieyi Gordon. The larger male para type from Lot b, in dorsal
aspect, left antenna and pereiopods II to V omitted.

sexually immature ; it also is markedly granulose, and the eye-stalk has a depression
distal to the bulbous base.

The male paratype from Lot b represented in Fig. 4 measures 4-60 by 4-72 mm.
It differs appreciably in outline from the female, since the carapace is relatively

ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS

narrow and the anterolateral spines project but a short distance beyond the orbit.
The lateral frontal spines as well as the anterolateral ones are longer and more
slender and the intercalated lobe on the orbital margin is less pronounced. The
areolation and the granulation of the dorsal surface of the carapace are also less
conspicuous. The chelipeds are equal, and the merus projects some distance beyond
the anterolateral spines (cf. Figs. 4 and ia). The spine at the inner angle of the
carpus is small ; the two pairs of granulose humps on the dorsal border of the palm

Fig. 5

. Sirpus zariquieyi Gordon. Carapace of male para type from Lot c, in dorsal

aspect.

tend to merge, and the fingers are, as in the female, rather longer than the dorsal
palmar margin. The walking legs are more slender than in the female ; the merus
of pereiopod V is rather more than three times as long as wide, and the ridges on
carpus, propodus and dactylus are less distinct. The details of epistome, buccal
cavity, orbit, external maxilliped, etc., are shown in Fig. 3 a. The abdomen is
narrowly triangular ; somites 3 to 5 are coalesced, and somite 7 is rather longer
than the preceding one (Fig. 2d). The first pleopod, represented in Fig. 6b, is a
long slender style reaching almost to the apex of the abdomen ; the second pleopod
is short, about one-fourth of the first.

ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS

The specimen received on loan from the Leiden Museum is also a male, and though
measuring only 3*42 by 3*58 mm. also has the first pleopod well formed and nearly
as long as the abdomen. It differs from the larger male chiefly in that the penulti-
mate anterolateral spine and the intercalated spine of the orbital border are each
only incipient (cf. Figs. 5 and 4). The separation of basal and distal portions of
the eyestalk is also more marked by the depression mentioned as present in the
females from Lot a. This specimen has, in addition to the two dark posterior reni-

Fig. 6. Trachycarcinus glaucus Ale. & And. a. First and second pleopods of male

syntype.

Sir pus zariquieyi Gordon, b. First and second pleopods of largest male paratype.

form areas, a similar dark patch anteriorly near each orbit. The well-formed pleo-
pods of all these minute males suggests that they may be capable of impregnating
females, though they are rather young.

Habitat. Dr. Zariquiey informs me that the specimens from Lot b " were
captured a short distance from the coast, at a depth of about twenty metres, amongst
long Cystoseira algae growing in great quantity on the rusty iron plates of the hulk
of a sunken vessel. On examining at home the pieces of rust and algae, I found the

52 ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS

Sirpus at the base of the said algae and on the pieces of rusty iron among other
small algae (not determined). In the same place I found hundreds of Catapagn-
roides timidus (Roux), Thoralus cranchii (Leach), Galathea bolivari Zariquiey, Por-
cellana longicornis (Pennant), Alpheus dentipes Guerin and Synalpheus laevimamts
(Heller)." ... " The exact point is next to Punta Oliguera facing a sector of
the coast called Los Cayals, in the vicinity of Cadaques (South of Cabo de Creus),
Provincia de Gerona, Spain."

Sirpus monodi n. sp.

Material, (a) Collected by M. R. Sourie, at Dakar, 7.^.1950. See also under
habitat p. 57. One ovigerous female, the holotype (c.l. = 7 mm., c.b. = 7-3 mm.).

(b) Collected by MM. Th. Monod and P. Budker on io.i.41 near Dakar, dredged
by the " Cabellou," station 6 bis (" a TEst de la bouee du Banc de Bel Air, 14 42' N.,
17 24' W.), in about 12 m. Bottom " Sable grossier." One male [c.l. = 4-9 mm.,
c.b. = 5*1 mm.).

(c) Collected by M. Paraiso at Goree, near Dakar, on i.ix.50. Entangled in net
used for fishing " langoustes " (Panulirus regius) and set in a few metres depth only.
One female with genital openings long narrow slits and so probably slightly immature
(c.l. = 3-9 mm., c.b. = 3*8 mm.).

Description of holotype. The carapace, represented in dorsal aspect in Fig.
7A, is only a trifle wider than long ; the dorsal surface is markedly nodular in such
a manner as to emphasize the areolation. The general outline is hexagonal, with
the anterior and posterior margins shorter than the other sides (the orbit being con-
sidered along with the anterolateral margin as one side of the hexagon). The
front comprises three subequal, bluntly triangular spines, the median one situated
below the level of the lateral ones and directed obliquely forwards and downwards.
The orbit is oblique, rather shallow but well formed anteriorly, imperfect postero-
ventrally as shown in Fig. 8a, so that the retracted eye is protected only by the
postocular spine and the corneal area is visible both dorsally and ventrally. The
main lobe of the dorsal orbital margin terminates anteriorly in a blunt supraorbital
spine or tooth, and is separated by a narrow hiatus from the blunt lobe intercalated
between it and the postocular (or first anterolateral) spine. The ventral margin of
the orbit is formed partly by segments 2 + 3 of the antennal peduncle and partly
by the lobe bearing the infraorbital spine (Fig. 8a, 2), behind which it gradually
fades out.

The anterolateral margin is divided into four spines, of which the third is the
smallest and the anterior is the longest ; the last, though short, is much wider
than the third. The posterolateral margins are straight or even a trifle convex,
and somewhat convergent. The submarginal crest parallel to the posterior margin
is low and interrupted medially.

The two protogastric lobes bear each a blunt, forwardly directed spine or tooth,
while the summit of the mesogastric lobe has a prominent wide cluster of granules.
On the summit of the cardiac lobe is a pair of granular clusters. A high blunt
hump, also beset with granules, is situated near the base of the last anterolateral
spine. On either side of the shallow depression separating mesogastric and cardiac

ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS

lobes is a prominent branchial lobe, with a prominent granular hump on its summit
and a smaller one some distance behind this. In addition, there are numerous
granular patches or striae on the posterior half of the carapace.

I mm.

Fig. 7. Sirpus monodi n. sp. Holotype, an ovigerous female, a. Carapace and left
cheliped, in dorsal aspect, b. Carpus and chela of right cheliped.

The antennae are long and setose, but the setae are simple, not pluomse. The
blunt spinose anteroventral angle of segments 2 + 3 of the peduncle is visible in
dorsal aspect (Figs, 7 a and 8a) ; segments 4 and 5 are of equal length, but the former
is the more robust.

ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS

The eyestalk is swollen and bulbous proximally, slender and cylindrical throughout
most of its length ; the distinctly pigmented corneal area is rather small and ter-
minal.

I mm.

Fig. 8. Sirpus monodi n. sp. Holotype. a. Right anterior portion of carapace, in
ventral aspect, to show details of orbit, epistome, buccal cavity, etc. b. Pereiopod V
and distal segments of pereiopod IV. I. Supraorbital spine. 2. Infra-orbital spine.

ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS 55

The chelipeds are equal ; the right, which is detached from the specimen, is repre-
sented at a higher magnification in Fig. 7B. Part of the merus is visible in dorsal
aspect beyond the two posterior anterolateral spines. The carpus and palm of the
chela are both conspicuously nodular ; the fingers are rather longer than the dorsal
margin of the palm, and there is a proximal tuft of setae on the upper border of the
dactylus, about one-third of the way along. When closed, the cutting edges of the
fingers are in contact distally, leaving a narrow proximal gap ; the outer surface of
each finger bears a low ridge. The maximum height of the palm is almost equal to
the length of the dorsal margin. The spine on the inner border of the carpus is
conspicuous.

The walking legs, pereiopods II to V, are bent inwards round the large bunch of
ova and so have been omitted from Fig. ja. The distal segments of pereiopod IV
are represented in Fig. 8b ; there is a low median dorsal ridge or carina on the
carpus, two dorsal ridges on the propodus and several longitudinal ridges on the
dactylus, which is not compressed. Pereiopod V is shown also in Fig. 8b ; the
merus is not quite three times as long as wide, the dorsal ridge on the carpus is more
prominent distally, the propodus is more compressed and the two dorsal ridges are
very close together.

The antennules are set in deep sockets beneath the front as represented in Fig.
8a ; they are folded obliquely, and there is a conspicuous crest or carina on the
swollen basal segment of the peduncle. The epistome sends forwards a narrow
prominent median carina to meet the backward projection of the front. The
anterior margin of the buccal cavity is distinctly marked by a ridge, as shown in
Fig. 8a.

All seven somites of the narrow abdomen are free ; the maximum width across
somites 3 and 4 scarcely exceeds the width of the posterior margin of the carapace.
Somites 5 and 7 are equal, and each a trifle longer than somite 6 ; the terminal one
is narrowly triangular.

The ova are numerous and small (a trifle larger than in the type species).

The male. The single male specimen is represented in dorsal aspect in Fig. 9.
Most of the pereiopods are detached from the left side, while the last one, pereiopod
V, is missing on the right side. The carapace is very similar in outline to that of the
female ; the frontal and anterolateral spines are all rather blunt and massive and
the supraocular lobe is broad and rounded. As the specimen is younger than the
holotype the areolation of the carapace is less pronounced, but the nodular promi-
nences on the principal lobes are distinct. The eyestalks are relatively more robust
than those of the holotype. The crest near the posterior margin of the carapace is
more distinct and continuous. The abdomen is rather damaged near its junction
with the carapace, as Dr. Monod had dissected out the pleopods from one side and
mounted them on a slide, so that my reconstruction of that part may not be quite
correct (Fig. 9). Somites 3 to 5 are coalesced ; the distal somites are very similar
to those of 5. zariquieyi, represented in Fig. 2D. The first pleopod is slightly dis-
torted in the micropreparation, but is a long slender style reaching nearly to the
apex of the abdomen. It is very similar to that of S. zariquieyi, as is the second
pleopod. The chelipeds are equal, rather more massive than those of the holotype,

5 6 ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS

although as yet the nodular condition is less highly developed on carpus and palm ;
the cutting edges of the fingers each bear six well-formed lobes or teeth, the proximal
ones being more pronounced than in the older female.

Remarks. The rather immature female is also very similar to the holotype.

Fig. 9. Sirpus monodi n. sp. Male paratype, in dorsal aspect, pereiopods on left side

omitted.

This species undoubtedly belongs to the genus Sirpus and, like the type species,
it reaches sexual maturity at a remarkably small size. 5. monodi differs from S.
zariquieyi chiefly in the following respects : (i) It is a much more nodular form, with
shorter and much more massive spines on front, orbits and anterolateral borders,
(ii) There is no marked difference between the sexes as regards the shape of the
carapace, (iii) The chelipeds and walking legs are relatively more robust, and the
latter rather shorter as well. The propodus of pereiopod V is rather more compressed,
(iv) The setae on the antennal flagellum are simple and rather short, instead of long

ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS 57

and distally plumose ; the segments of the peduncle are much more robust (cf.
Figs. 8 A and 3A).

Habitat. Dr. Monod informs me that the holotype was collected " a maree
basse, a Dakar (Anse Bernard) au niveau de Tarrleurement de tufs volcaniques (la
meme station a fourni des Perimela 1 denticulata) . M. Sourie me dit que le petit
crabe, tres homochrome, etais tres difficile a apercevoir sur le sable." Summing
up for all three localities he writes : " TEspece semble done, pour le moment etre
littorale, depuis la zone intercotidale jusqu'a 10-15 m." (letter dated 11. hi. 53).

The holotype will be deposited in the Paris Museum, the male paratype in the
I FAN at Dakar, and the female paratype in the British Museum Collection.

Since the MS. went to press, Dr. Monod has found an ovigerous female measuring
4*5 by 4*8 mm. amongst some undetermined material in the Paris Museum. The
specimen was collected by A. Gruvel on 2.iv.o8, at Pointe Cansado, C. Blanco,
Morocco, much further north than the type material. It agrees well with the
holotype but, being smaller, it has many more granular st nations on the dorsal
surface of the carapace.

THE SYSTEMATIC POSITION OF THE GENUS

When Dr. Zariquiey first sent me the two small female specimens comprising
Lot a of 5. zariquieyi he referred to them in his letter dated 16.ix.47 as " dos Maiidae
formas juveniles de sp. ? ". On examining them my first impression was that they
might be very young stages of Pirimela denticulata (Montagu) but this proved not
to be the case, and I also was inclined to regard them as young stages of a species
of Spider Crab. Now that an ovigerous female is available and I have examined
them more carefully, I am convinced that they are not referable to any family of
the large sub tribe Oxyrhyncha. No Oxyrhynchous crab has the median frontal
spine so well formed that the front looks trispinose ; when present the median
spine or rostrum is minute, and much more ventral in position than the well-de-
veloped rostral (or pseudorostral) horns (e.g., in Maia or in Mithrax). Moreover, in
the Oxyrhyncha it is unusual for the anterolateral and posterolateral borders of the
carapace to be clearly separated, and the basal segment (or more correctly segments
2+3) of the antennal peduncle usually forms a larger part of the lower orbital
border. The general form of the carapace in the ovigerous female is distinctly
Cancroid, in the wide sense employed by Rathbun, 1930, apart from the front, the
median spine of which slopes obliquely downwards and so lies in a different plane
from the other two (Figs, ia and 2b). It would appear that my first impression —
that they are closely related to Pirimela — was fairly sound. Dr. Balss, who has
done so much to elucidate the classification of the Oxyrhyncha and to whom I sent
some sketches, writes (6.x. 52) : " Das Bild einer jugendlichen Krabbe, das Sie mir
iibersandt haben, gibt nicht eine Oxyrhynche Krabbe, sondern eine primitive Por-
tunide (Carcinus maenas ? oder eine verwandte Art)." He was relying entirely
on memory, being too ill to go to Munich to consult the literature. Dr. Monod, to

1 According to Monod (1933, Bull. Com. d'Etudes hist. sci. Afr. Occ. Fr. 15, 2-3 : 52) Pirimela is
etymologically incorrect.

5«

ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS

whom I also sent sketches and who has since found the specimens belonging to the
second species 5. monodi, writes (9.xii.52) : " ne devrait-on pas chercher du cote
des Atelecyclides, aux alentours de T r achy car cinus par example ? " Thus all three
of us are agreed that the genus belongs to the Cancroid or Cyclometopous Crabs in
the wide sense used by Rathbun, 1930.

Car cinus maenas or Carcinides maenas (L.) is one of the few species of European
brachyura the postlarval development of which is known in detail (see, e.g., Shen,
1935, p. 19, text-fig. 20), and the specimens from Cadaques do not fit into this
series. The nearest approach to these specimens that I have discovered so far is a

Fig. 10. Two figures after Cano, at twice his magnification.

a. " Maja verrucosa, stadio postlarvale " from 1893, taf. 34, fig. 29 (x 8).

b. " Fase adulta (giovane) di Pirimela " from 1891, 1892, tav. iv, fig. 4G (Cano's
magnification not specified).

young crab stage referred to Maia verrucosa H. M.-Edw. by Cano (1893, Taf. 34,
fig. 29), and which I have illustrated ( X 2) in Fig. ioa. The size of this specimen
is approximately 4x4 mm., but I do not think that it can be the first, second or
later young crab stage of M. verrucosa — Cano did not rear it from the megalopa, of
course. Miss Lebour seems to have accepted Cano's interpretation of his specimens,
so it is perhaps presumptuous of me to question this stage. The development of
Maia squinado (Herbst) has been described by Lebour (1927), who obtained the
prezoea, first and second zoeae and megalopa from the eggs. But, as she states on
p. 797, in no case did the megalopa turn into a crab. The megalopa of Maia squinado

ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS 59

from which she obtained the first three young crab stages was obtained from the
plankton. What strikes one about these stages is their marked resemblance to
young Portunids ; although the front is rather more pointed medially, the general
resemblance of these young Maia to the corresponding stages of several species of
Portunus and of Carcinides is unmistakable (cf. Lebour, 1927, PI. Ill, fig. 10 with
Lebour, 1928, PI. VI, figs. 1-3 and pi. 7, figs. 1 and 3). The young crab stages of
Macropodia and of Inachus, on the other hand, are unmistakably Oxyrhynchous
(see Lebour, 1927, PL 3, figs. 5 and 8). I cannot help wondering whether the mega-
lopa from which these three young stages were obtained is really referable to the
genus Maia.

To return to Cano's young "Maja verrucosa," it resembles the specimens from
Cadaques as regards the front, eyestalks and orbits, and in the number of antero-
lateral spines. If the supraorbital spines were omitted from Figs. 4 and 5 and the
orbital margins were carried straight back from the bases of the lateral frontal spines,
and the anterolateral spines on either side of the smallest one were exaggerated
slightly, the result would be much as in Cano's figure (Fig. ioa).

Bouvier (1942) has separated the crabs of " la tribu des ' Corystoidea ' " from the
rest of the Brachyrhyncha ; to the Corystoidea he has referred the Euryalidae
(= Corystidae restr.), Atelecyclidae and Cancridae of Rathbun (1930, pp. 10, 148
and 176) and others, placing the genera in five families (see pp. 47-48 for his scheme).
The genus Trachycarcinus he places in the family Corystidae, subfamily Atele-
cyclinae ; Pirimela in the monotypic family Pirimelidae. I had noted the rather
striking resemblance between the males of Sirpus zariquieyi and Trachycarcinus
corallinus Faxon for example. In the British Museum Collection the genus is repre-
sented by one male syntype of T. glaucus Alcock & Anderson and one male specimen
of T. balssi Rathbun received from Dr. Sakai of Japan. Sirpus agrees with the
diagnosis of Trachycarcinus given by Rathbun (1930, p. 164) as regards the orbits,
which are " large, with forward aspect, imperfect," but it differs in many respects :
(i) The eyestalks are retractile within the orbits but they are not " very small,"
nor are the eyes " dull and faintly pigmented." (ii) The front is wider and, though
trispinose, the spines do not lie in one plane as they do in Trachycarcinus. (iii) The
anterior margin of the buccal cavity is very distinct and the ridges of the endostome
are less pronounced than in T. balssi for example. The external maxillipeds also
differ in the two genera ; in Trachycarcinus there is a considerable gap between the
maxillipeds anteriorly, the anterior inner angle of the ischium is advanced and the
merus is " obliquely truncated without emargination at antero-internal angle "
(Rathbun, 1930, p. 164, pi. 72, fig. 6)). (iv) The carapace is not pentagonal with
long, nearly straight, anterolateral margins, but hexagonal (more ovoid in the female
of the type species 5. zariquieyi). (v) The antennules do not fold longitudinally or
lengthwise as in the Atelecyclidae, but obliquely, as shown in Figs. 3A and 8a. (vi)
The abdomen of the male is more narrowly triangular, with a longer apical somite
and somites 3 to 5 completely coalesced (cf. Figs. 2D with 2A — somites 4 and 5 are
imperfectly coalesced in Trachycarcinus glaucus and in T. corallinus but all seven
somites are free in T. spinulifer) ; and, judging from the figure given by Rathbun
(1930, pi. 72, fig. 5), that of the female is also narrower, with a much longer apical

ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS

somite, in Sirpus. (vii) The pleopods of the male also differ markedly in the two
genera as represented in Fig. 6a and b ; in Trachycarcinus the first pleopod is rela-
tively straighter, more robust and shorter, while the second pleopod is longer than
the first and has a terminal lash. In all the genera of the Atelecyclinae that I have
examined, namely Telmessus, Peltarion, Atelecyclus} Trachycarcinus and Pliosoma,
pleopod 2 of the male is as long as or longer than pleopod i except in Pliosoma ; in
the latter pleopod 2 is very short, measuring only about one-sixth of pleopod i,
and the genus may be referable to a different subfamily. I should expect Erimacrus
and Trichopeltarion to conform to the general pattern (i.e., to have pleopod 2 long),
but I have no material at my disposal ; Rathbun and Bouvier both omitted to

Fig. ii. Pirimela denticulata (Montagu). A young male from Jersey (width of carapace
approx. io*7 mm.), a. Part of carapace, in ventral aspect, to show details of orbit,
epistome, buccal cavity and third maxilliped, etc. b. First and second pleopods, with
apex of each further enlarged, of young male (10-4 x 11-4 mm.).

mention the form of the male pleopods. In the Corystidae (or Corystinae of Bouvier)
pleopod 2 varies in relative length, being longer than pleopod 1 in Pseudocorystes
and Podocactes, about the same length as pleopod 1 in Gorneza and Jonas, shorter
than, but exceeding half the length of, pleopod 1 in Corystes. In the family Bellidae
of Bouvier's scheme (Acanthocyclinae in that of Rathbun) pleopod 2 reaches nearly
to the apex of pleopod 1 in Bellia, but is rather less than half as long as pleopod 1
in Acanthocyclus. In the genus Cancer pleopod 2 is rather longer than pleopod 1 —

1 Stephensen (i945» P- 222) is in error when he says of the subfamily Atelecyclinae, " Pip. 2 short,
of usual shape " ; Brocchi, to whom he refers, says on p. 103 that pleopod 2 of Atelecyclus cruentatus
Desm. is long, and his fig. 199 shows a long slender appendage which exceeds pleopod 1 (figs. 197-8).

ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS

family Cancridae of Bouvier's scheme — whereas it is short in Thia, Kraussia (family
Thiidae) and in Pirimela (family Pirimelidae) (see Fig. iib).

Fig. 12 represents a small male specimen of Pirimela denticulata (Montagu) from
Jersey in dorsal aspect. The areolation of the carapace is very similar to that found
in both species of Sir pus. The front is trilobed, the median one being the most
prominent, and all lie in the same plane. The antennae are rather long but as a
rule are not setose, though a few very short plumose setae may be present in some

Fig. 12. Pirimela denticulata (Montagu). Young male, in dorsal aspect, left pereiopods

omitted.

specimens. The orbit is closed and narrow, but the dorsal border comprises three
lobes, separated by two notches ; the eystalks are rather short, but the basal portion
is bulbous (Fig. iia). The anterolateral margin bears five forwardly directed
spinose lobes. Pirimela denticulata grows to a larger size and is far less precocious
than either of the species of Sirpus — the first trace of the genital opening appears
when the females measure 9 to 10 mm. in maximum width, and the opening is not
fully formed until they measure 11 to 12 mm. across the last pair of anterolateral
spines. It is not so easy to ascertain when the males reach sexual maturity ; in a

ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS

small specimen measuring 7 mm. in width the first pair of pleopods is short and
obviously immature, but by the time the carapace measures 8-5 to 9 mm. in width
these pleopods are well-developed long styles. Sirpus appears more closely related
to Pirimela than to Trachycarcinus in the following respects : (i) The manner in

Fig. 13. a. Portumnus biguttatus Risso. b. Portumnus latipes (Pennant). Anterior
portion of carapace, in ventral aspect, to show details of front, orbit and eyestalk,
antenna and antennules, epistome and anterior margin of buccal cavity.

which the antennules are folded, the distinct anterior margin to the buccal cavity,
the well-pigment ed eyes and the form of the external maxilliped (cf. Figs. 3 A and
8a with iia). (ii) The form of the abdomen in both sexes, (hi) The form and rela-
tive lengths of the pleopods of the male (cf. Figs. 6b and iib). (iv) The lobulation
of the carapace — the female of S. zariquieyi also approaches more nearly to Pirimela
in general form of the carapace (cf. Figs, ia, 7A and 12). (v) The walking legs are
also very similar. At present we do not know how large either of the species of

ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS 63

Sirpus may become, but they would appear to be really of very small size, since even
in the smallest males the pleopods are already well formed, and in all but the smallest
female the genital openings are visible. It is unlikely that the females would alter
appreciably with increase in size, since in each species the largest bears ova.

Cano (1891, tav. iv, fig. 4G — published in 1892) gives a figure of what he calls the
" fasa adulta (giovane) di Pirimela," which I have reproduced at twice his magnifi-
cation in Fig. iob. As Cano does not appear to give his magnification I do not
know the size of his specimen. It differs appreciably from all the stages that I have
examined chiefly in having wide shallow orbits, long eyestalks and only four antero-
lateral spines — in these respects his specimen recalls 5. zariquieyi even more than
Pirimela, although the front is nearer to that of the latter (Cano's specimen is repre-
sented in his figure as rather asymmetrical). At present the postlarval develop-
ment of Pirimela denticulata is not known ; it seems unlikely that Cano's specimen
is a young crab stage of Pirimela, but if it is the orbit must alter markedly with age
and the number of anterolateral spines increase to five (cf. Figs. 12 and ioa).

The wide shallow orbit alone would not necessarily exclude these new species
from Cadaques and West Africa from the genus Pirimela. The two European
species of the genus Portumnus differ from each other markedly as regards orbit and
eyes talk as represented in Fig. 13 A and b. In Portumnus biguttatus 1 Risso the orbit
is closed, narrow, and deep enough to conceal the cornea of the retracted eye dor-
sally, though not ventrally (owing to the deep gap or hiatus in the ventral margin) .
The trilobed front is separated by a wide shallow emargination from the low rounded
supraorbital lobe. The principal segment (2 + 3) of the antennal peduncle fills the
gap between the orbit and the antennular fossa, and, in the specimen from which
the figure was made, is slightly movable. In Portumnus latipes (Pennant) the front
is more markedly three-pronged ; the orbit is wide and shallow, affording no con-
cealment for the cornea of the retracted eye ; moreover, there is a wide gap between
the infraorbital spine or angle and the antennal peduncle (Fig. 13B). Sirpus and
Pirimela differ from each other as regards orbit and eyestalk in much the same way
as do these two Portumnus species. However, Sirpus differs sufficiently from Piri-
mela as regards the form of the front, the number and form of the anterolateral
spines, the more setose antennae, and in the absence of a " lobe portunien " on the
endopodite of the first maxilliped (see Bouvier, 1942, p. 32, fig. 15), to justify the
erection of a new genus. But these differences hardly seem sufficient to justify a
new family for the reception of Sirpus.

Authors do not seem to agree as to the systematic position of the genus Pirimela.
Pest a (191 8, pp. 386-387) places it with Cancer and Carcinides in the family Can-
cridae ; most authors place Carcinides in the Portunidae. Nobre (1931, p. 84 ;
1936, pp. 48-49) refers it to the Cancridae along with Cancer and Xantho ! In the
Plymouth Marine Fauna, 2nd edition, 1931 the family Cancridae includes Cancer
and Atelecyclus, while Pirimela is referred to the monotypic family Pirimelidae.
Bouvier (1942, pp. 47-48) includes the Cancridae and the Pirimelidae, each with
only the genus from which the family name is derived, in his Corystoidea, which
he removes from the Brachyrhyncha (and yet he refers to Pirimela as having a

1 Referred to the genus Portumnoides by Bohn 1902, p. 448.

64 ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS

narrow " lobe portunien " as mentioned above). Lebour (1944, p. 15) writes : "As
far as we know from its larval stages Pirimela is probably related to Portunus and
Cancer, perhaps more nearly to Cancer." Enough has been said to show that the
classification of the Brachyura is in need of thorough revision. In the meantime
I am inclined to place Sir pus near to Pirimela rather than to Trachycarcinus, although
it shows some resemblance to the latter. In the shape of the abdomen in both
sexes and also of the male pleopod 1 it is much nearer to Pirimela than to Carcinides
and the dactylus of pereiopod V is not compressed as in the latter genus, which is a
primitive Portunid (see p. 57 for Dr. Balss' view).

The two genera of the family Pirimelidae may be distinguished as follows :

a. Front trilobed, the lobes in one plane ; orbits narrow and complete ; antero-

lateral margin divided into five teeth ; antennal flagellum not obviously-
setose ; specimens reaching a fair size (up to 25-30 mm. across carapace)

Pirimela, Leach.

b. Front trispinose, the spines not in one plane ; orbits wide and incomplete ventro-

posteriorly ; anterolateral margin divided into four teeth or spines of which
the third is the smallest ; antennal flagellum setose ; specimens of very small
size ........... Sirpus, Gordon

The study of these small crabs was particularly difficult, I think, because of their
extreme sexual precocity, as a result of which they have inevitably retained a number
of juvenile characters. For example, the great fronto-orbital width, the projecting
front, the relatively large frontal, orbital and anterolateral spines and the long setose
antennae are all strongly reminiscent of the first postlarval stage of Cancer pagurus,
figured by Cunningham (1898, Proc. zool. Soc. London, p. 204, pi. xxi, fig. 1). Shen
has figured the outline of the carapace of the first nine young crab stages of Carcinus
maenas (1935, p. 19, fig. 20), so that one can see at a glance how the shape and
relative proportions vary with age. As regards size, the largest ovigerous females
of the genus Sirpus would be equivalent to the sixth young crab stage of Carcinus
and sexual maturity is reached even at a smaller size. The two species of the genus
Sirpus may therefore be regarded as neotenous. de Beer (1951, 'Embryos and
Ancestors,' revised edition, p. 52), applies the term neoteny to "cases where the
adult animal retains larval characters " ; where a complicated metamorphosis
occurs, however, it is more likely to be the postlarval characters that are retained
by the adult.

A cknowledgments .

I wish to thank Dr. Zariquiey of Barcelona, Dr. Boschma of Leiden and Dr. Monod
of Dakar for sending me these very interesting specimens for study. I am also
grateful to Dr. Balss of Munich and Dr. Monod for giving me their views on the
relationship of the genus, and to my colleague Mr. A. C. Townsend for assistance in
finding a name for it.

The authorities of the British Museum (Nat. Hist.) are much indebted to Dr.
Zariquiey for presenting to their Collection the holotype and a male paratype of 5.
zariquiey i, and to Dr. Monod for presenting the female paratype of 5. monodi.

ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS 65

LIST OF WORKS CITED

Bohn, G. Des mecanismes respiratoires chez les Crustaces Decapodes. Bull. sci. Fr. Belg.,

36 : 178-351, 209 text-figs.
Bouvier, E. L. 1942. Les Crabes de la Tribu des " Corystoidea." Mem. Acad. Sci. Inst.

France, 65, 1941 [1942] : 1-52, 18 text-figs.
Cano, G. 1891 [1892]. Sviluppo postembryonal dei Cancridi. Boll. Soc. ent. ital. 23 :

146-158, pis. 3 and 4.

1893. Sviluppo e Morfologia degli Oxyrhynchi. Mitt. zool. Sta. Neapel. 10 (4) : 527-

583. Pis.
Gordon, I. 1953. On a new Crab from Cadaques on the north-east Coast of Spain (Sirpus

zariquieyi n.g. and sp.). Eos, Madrid, 28 (4) : 308-314, 5 text-figs.
Lebour, M. V. 1927. Studies of the Plymouth Brachyura. I : The Rearing of Crabs in

Captivity, with a description of the Larval Stages of Inachus dorsettensis , Macropodia

longirostris and Maia squinado. J. Mar. biol. Ass. U.K., n.s. 14 (3) : 795-814, pis. 1-4.

1928. The larval Stages of the Plymouth Brachyura. Proc. zool. Soc. Lond., 1928 :

473-56o, pis. 1-16.

1944. The larval stages of Portumnus (Crustacea Brachyura) with notes on some other

genera. /. Mar. biol. Ass. U.K., n.s. 26 (1) : 7-15, 5 text-figs.

Nobre, A. 1 93 1. Crustaceos Decapodes e Stomat6podes marinhos de Portugal. Inst,
zool. Univ. Porto, 1-307, 2 pis., 144 text-figs.

1936. Same title. Vol. IV in Fauna Marhina de Portugal, 1-216, 61 pis.

Pesta, O. 191 8. Die Decapodenfauna der Adria. Versuch einer Monographic : x + 500,

1 map, 150 text-figs. 8°. Leipzig & Wien.
Plymouth Marine Fauna. 2nd ed., 1931. Mar. biol. Ass. U.K. : 1-371 (Crustacea, pp.

150-222).
Rathburn, M. J. 1930. The Cancroid Crabs of America of the families Euryalidae, Portu-

nidae, Atelecyclidae, Cancridae and Xanthidae. Bull. U.S. Nat. Mus. 152 : 1-609, 230

pis., 85 text-figs.
Shen, C. J. 1935. An investigation of the post-larval development of the Shore-Crab Car-

cinus maenas, with special reference to the external secondary sexual characters. Proc.

zool. Soc. Lond., 1935 (1) : 1-33, 28 text-figs.
Stephensen, K. 1945. The Brachyura of the Iranian Gulf. Danish Scientific Investigations

in Iran, Pt. IV : 57-237, 60 text-figs.

PRESENTED

1 NOV 1953

PRINTED IN GREAT BRITAIN BV
ADLARD AND SON, LIMITED,
BARTHOLOMEW PRESS, DORKING

6 FEB 1954

MACROTRITOPUS
PROBLEM

W. J. REES

BULLETIN OF
THE BRITISH MUSEUM (NATURAL HISTORY)
ZOOLOGY Vol. 2 No. 4

LONDON: 1954

THE MACROTRITOPUS PROBLEM

W. J. REES, D.Sc.

Pp. 67-100 ; PL 3 ; 17 Text-figures

BULLETIN OF
THE BRITISH MUSEUM (NATURAL HISTORY)
ZOOLOGY Vol. 2 N0.4

LONDON: 1954

THE BULLETIN OF THE BRITISH MUSEUM
(NATURAL HISTORY), instituted in 1949, is
issued in five series corresponding to the Departments
of the Museum, and an Historical Series.

Parts appear at irregular intervals as they become
ready. Volumes will contain about three or four
hundred pages, and will not necessarily be compiled
within one calendar year.

This paper is Vol. 2, No. 4 of the Zoological
series.

PRINTED BY ORDER OF THE TRUSTEES OF
THE BRITISH MUSEUM

Issued January, 1954. Price Twelve Shillings.

THE MACROTRITOPUS PROBLEM

By W. J. REES, D.Sc.

SYNOPSIS

The Atlantic species of the larval octopod genus Macrotritopus (M. equivocus, M. scorpio, M.
hempi and M. danae), together with some Macrotritopus-like larvae described by Degner (1925),
have been re-examined, and the type material of all but the first two species has been available
for study. All these forms were found to be growth stages of Scaeurgus unicirrhus (Delle
Chiaje), an octopus of the continental shelf. The Macrotritopus larvae appear to be able to
delay settlement if they are swept over very deep water by water movements, and to continue
growing to nearly twice the normal size for metamorphosis. These large larvae (found also in
other groups of animals) are now termed extended pelagic stages, and their significance in dis-
persal and in maintaining the homogeneity of populations of widely distributed species is
discussed.

CONTENTS

a. Introduction .....

b. The Genus Macrotritopus .

(a) Octopus (Macrotritopus) equivocus Robson

(b) Octopus (Macrotritopus) scorpio (Berry)

(d) Octopus (Macrotritopus) danae Joubin & Robson

(e) Scaeurgus (Troschel sp. juv.) Degner.

c. A Re-examination of Macrotritopus Larvae

(a) Macrotritopus danae Joubin & Robson

(b) Macrotritopus kempi Robson and other " Discovery

(d) Specimens from Biscayne Bay, Florida

d. A Comparison of the Species

(a) Mantle length

(b) Head width

(d) Sucker diameter

(e) The radula

(f ) Gill filaments

e. The Breeding Season

(a) The Mediterranean

(b) The Atlantic

f. The Significance of the Macrotritopus Larva

g. Identity of Macrotritopus with Scaeurgus
h. Conclusion ......

1. References ......

zool. 11, 4.

Page

larvae

70
70

72
73
74

74
76

77
78
81
82
83
83
86

87
91

9i
91
92

94
95

97
98

70 THE MACROTRITOPUS PROBLEM

a. INTRODUCTION

This study originated from a desire to carry further earlier work on the planktonic
stages of Octopus vulgaris (Rees, 1950, 1952). In particular, it was hoped that late
planktonic stages of this species could be obtained from a search of plankton hauls
made in the tropical Atlantic by the ships of the " Discovery" Investigations. During
the course of this work a number of Macrotritopus larvae were collected, and it was
the remarkable similarity of the youngest Macrotritopus larvae to those of Octopus
vulgaris (except in the very long third arms) which led me to re-examine them in
relation to a known adult or adults.

The startling agreement of the chromatophore patterns in Macrotritopus and
Octopus vulgaris suggested some relationship, but further investigations showed con-
clusively the Macrotritopus could not be linked with any littoral species. Final
conclusions indicate that all the so-called Macrotritopus spp. of the Atlantic belong
to one species, and that the adult is Scaeurgus unicirrhus (Delle Chiaje), a benthic
octopod living at moderate depths near the edge of the continental shelf.

b. THE GENUS Macrotritopus

Grimpe (1922) proposed the name Macrotritopus as a generic name for a juvenile
pelagic octopod, Octopus gracilis Verrill, because the third arms greatly exceeded the
others in length. Later, Robson (1929&, p. 168, et seq.) changed Verrill's name to
equivocus because the name gracilis was preoccupied. While recognizing that Macro-
tritopus was only a convenient label for juveniles, whose adult stages might or might
not be known, he referred four other forms to it, viz. :

Octopus scorpio (Berry), Macrotritopus kempt Robson, Octopus bandensis Hoyle
and Octopus elegans Brock. Of these, Octopus (Macrotritopus) equivocus Robson and
the first three above were described from the Atlantic and the Mediterranean. The
last two are Indo-Pacific species and need not concern us further here. Later,
Joubin & Robson (1929) described Macrotritopus danae from the collections of the
Danish Dana Expeditions. All the known Atlantic specimens of Macrotritopus are
juveniles exhibiting no sings of hectocotylization.

It has also been found necessary to reconsider some larvae described by Degner
(1925) as Scaeurgus (Troschel sp. juv.), because on re-examination they prove to
be Macrotritopus larvae.

The Macrotritopus forms which have been described are so distinctive in appearance
that their true identity is masked by larval characteristics. In this connection the
views of earlier workers are of interest. Verrill (1884), in considering his own
Octopus gracilis (i.e., equivoca), was " very certain it is not the young of any known
species." Robson (1929&) was unable to make up his mind about the status of these
forms, as is evident from his statement : " We must either conclude that they are
the young forms of species hitherto undescribed or that they undergo some kind of
metamorphosis in the course of subsequent development, as a result of which they
assume the adult form of some described species." More recently, Pickford (1945)
in discussing the American forms M. equivoca and M. scorpio, expressed the opinion
that " There is no possibility that these strikingly distinct young animals with their

THE MACROTRITOPUS PROBLEM

very long third arms could belong to a vulgaris-\ike species." The same author
(1947) added " whatever they may be they are clearly not the young of any littoral
species." Voss (195 1) considered the possibility of M. scorpio being the larval form
of Scaeurgus but had insufficient material to reach any decision.

These views reflect the need for clearing up the status of the Atlantic forms of
Macrotritopus, and except for M. equivoca I have been privileged to re-examine all
the original material reported on by Degner (1925), Robson (1929), Joubin & Robson
(1929) and Voss (1951) together with many additional specimens from the plankton
catches of the " Discovery " Investigations, 1925-1939. I have been able to do this
through the generosity of Dr. Anton Brunn (for the loan of Dana specimens) ; Dr.
Gilbert Voss (for the use of two specimens of Macrotritopus from Florida) ; and Dr.
N. A. Mackintosh and Dr. Helen Bargmann, for "Discovery" material. I also
wish to thank Dr. G. R. de Beer, F.R.S., and Dr. H. W. Parker for valuable sugges-
tions, Mr. G. L. Wilkins for his superb drawings (PI. 3 and figures 1-6), and Mr. H. O.
Ricketts for his great care in examining large samples of plankton for octopod larvae.

(a) Octopus (Macrotritopus) equivocus Robson

Octopus gracilis Verrill 1884, p. 236 non Eydoux & Souleyet 1852, p. 13.

Macrotritopus equivocus Robson 1929a, p. 311.

Octopus {Macrotritopus) equivocus, Robson, 19296, p. 1G9.

Octopus {Macrotritopus) equivoca, Joubin & Robson 1929, p. 93.

Type locality. "Albatross" St. 2084, 40 16' 50" N., 65 05' 15" W., south of
Cape Sable, Nova Scotia, 1,290 fms. ; 1 specimen.

Table I
Measurements in mm.

Length of body (? ventral mantle length)

Mantle width

Interocular width ....
Length of head and body
Arm length :

1st

2nd ......

3rd

4th (damaged) ....

11 -o

7-0

6-5
17*0

19-0
21 -o
42*0

Verrill thought that the specimen was a juvenile of a species growing to a large
size and was convinced that it was not the young of any known species. He noted
the extraordinarily long third arms, the second arms being only half their length.
On the evidence available from the description, it agrees with other Macrotritopus
larvae noted later in having (in alcohol) large purplish chromatophores on the body
and the head. There are also purplish spots (i.e., chromatophores) in front of and
behind each sucker, this feature being common to all Macrotritopus and to many
other octopod larvae.

Verrill notes that the first 3-5 suckers on the lower pairs of arms are uniserial and

72 THE MACROTRITOPUS PROBLEM

that this arrangement is not found on the dorsal arms. This is unusual, but a minor
point, as the arrangement may be due to unequal contraction of the arms during
fixation.

Three features about this larva are noteworthy. It is the largest known specimen
of Macrotritopus ; it was caught over deep water and its occurrence (off Nova
Scotia) is well to the north of any other records of larvae of this genus.

(b) Octopus (Macrotritopus) scorpio (Berry)

Polypus scorpio Berry 1920, p. 299, pi. 16, fig. 4.
Octopus {Macrotritopus) scorpio, Robson 19296, p. 169.
Octopus {Macrotritopus) scorpio, Joubin & Robson, 1929, p. 93.

Locality. "Bache" St. 10204, off Biscayne Bay, Florida ; 20.iii.1914 ; 75-0 m. ;
1 specimen.

Berry gives total length as 22 mm., and 4-5 mm. as dorsal mantle length. The
measurements given below have been calculated from his figure in order to provide
some basis for comparison with other larvae. 1

Table II

Measurements in mm.

Ventral mantle length .

• 4'73

Dorsal mantle length .

• 5-89

Mantle width .......

• 3-85

Interocular width ......

• 3'8

Arm length :

ISt

• 3'i

2nd ........

• 7-27

3 r d

■ 14*9

4th

. 4-62

Mantle-arm index ......

. 31.8

Berry had some doubts about his species being specifically distinct from VerriU's
gracilis, but mentioned the inequality in the length of the first and second arms, the
still greater length of the third arms and the minute papulations as distinguishing
features. The chromatophore pattern is only briefly mentioned.

Macrotritopus kempi Robson, 1929a, p. 311.

Octopus {Macrotritopus) kempi, Robson, 1929&, p. 170.

Type locality. " Discovery " St. 276, 5 54' S., n° 19' E., off the mouth of the
Congo; 5.viii.i927; N 70B, 110-0 metres; 2 syntypes (B.M. 1947. 3. 12. 1-2).

Robson regarded this species as a very distinct one on account of its narrow head,

1 Berry's figure is not very satisfactory ; the head is said to be " distinctly narrower than the body,"
but as pointed out by Robson it is nearly as wide as the body in the figure; Berry's " dorsal mantle
length " of 4.5 mm. is probably the distance from the apex to the nuchal ridge, and the measurement
given in the table is the distance between the apex and the point midway between the eyes.

THE MACROTRITOPUS PROBLEM 73

wide mantle, the smooth surface of the skin, the size of the web and the distinctive
colour pattern. He also drew attention to the proportions of the third arm, and
mentioned that there were 11-12 filaments in each demibranch.

Robson (19296, p. 171) drew attention to the position where this species was
found over deep water and about 10 miles off the edge of the continental shelf. This
species is further discussed and the specimens re-described on pp. 76-77.

(d) Octopus (Macrotritopus) danae Joubin & Robson

Octopus (Macrotritopus) danae, Joubin & Robson, 1929, p. 87, fig. 1.
Octopus (Macrotritopus) danae, Joubin 1937, p. 33, fig. 33.

Type locality. "Dana" St. 1152, 30 17' N., 20 44' W., 23.ix.21 (160 metres).
Paratypes from the Western Mediterranean, off Guiana and off W. Cuba.

Joubin and Robson were clearly puzzled about the relationship of their species
with previously described forms. However, they created a new species chiefly on
the grounds that although the third arms were very long (74-80% of the total
length) , the second arms were also particularly well developed ; the third arm exceed-
ing the second by 1-7-2-1 times its length. Other differences were mentioned.

The measurements given by Joubin & Robson are given below for seven specimens
examined by them.

Table III

Measurements in mm.

No. of specimen.

2 134567

Mantle length (? dorsal) . . 13 I 3'5 IO 10-5 10 9 12

Arms : Left Right L. R. L. R. L. R. L. R. L. R. L. R.

1st • • . . 20 21 16 16 12 14 13 14 13 14 10 — ■ 13 12

28 32 25 25 22 20 22 21 22 23 17 18 21 19

48 37 38 36 37 4 2 37 33 35 35 28 32 39 36

27 29 20 — 18 18 18 19 17 19 15 12 19 17
Indices :

Mantle width, percentage length .57 48 60 57 55 61 —

Head width, percentage length . 46 44 50 52 50 55 50

3rd arms, percentage total length . 78 74 80 78 77 78 76

3rd arm > 2nd arm . . f 1 • 7 1*5 2-1 1 ■ 6 1-5 1-7 1 • 8

\ times

The above figures have not been used in my investigations, and all the specimens
which I have been able to see have been measured again as it was desirable to get
additional readings.

The " Dana " material forwarded to me, together with paratypes in the British
Museum (Natural History), contains a few specimens apparently not seen by Joubin
& Robson, so it has been deemed desirable to examine this material in detail in the
next section.

74 THE MACROTRITOPUS PROBLEM

(e) Scaeurgus (Troschel sp. juv.) Degner, 1925

" Thor " St. 184 : 38 10' N., 22 23' E., Gulf of Corinth ; 17.viii.1910 ; 65 metres
of wire.

Some octopod larvae from the Mediterranean in which the third arms were par-
ticularly well developed and long were described by Degner (1925, p. 79) ; these arc
essentially Macrotritopus-like forms which have to be considered in any discussion
of the Macrotritopus problem. Degner assigned these larvae to Scaeurgus because
one of his specimens was showing the beginning of a hectocotylus on the third left
arm. A sinistral hectocotylus is a characteristic of Scaeurgus and also of Pter octopus
tetracirrus, another Mediterranean species ; the latter species was not discussed by
Degner in relation to his larvae.

The relationship of these forms to other described species of Macrotritopus has
not been fully realized by earlier workers, and it is possible that they have been
misled by Degner's fig. 52, which portrays a larva quite unlike the usual Macrotri-
topus. It has been possible to re-examine these larvae in this investigation (see p. 74).

c. A RE-EXAMINATION OF Macrotritopus LARVAE

All the material of these larvae which it has been possible to gather together has
been re-examined. It includes the type material of Macrotritopus danae, M. kempi,
Degner's Scaeurgus, two new specimens from Florida (the type locality of M. scorpio),
as well as some additional specimens brought to light from an examination of hauls
made in the tropical Atlantic by the "Discovery" Expeditions (1925-1939).

(a) Macrotritopus danae Joubin & Robson

Larvae have been available from the following stations of the Danish " Dana "
Expeditions 1920-22 (see Schmidt, 1929) :

St. ii23 vii , 37 48' N., 2° 44' E., Western Mediterranean, 26.ix.21, S. 200,
200 metres of wire ; 1 specimen (8-25 mm. in ventral mantle length).

St. 1 1 24", 37° 15' N., 2 1 55' E., Western Mediterranean 27.ix.21, S. 200,
200 metres of wire ; 2 specimens (2-5, 675 and 9-0 mm. in ventral mantle length);
a'so Joubin & Robson's No. 2 specimen.

St. H74 iU , 5° 35' N., 51 08' W., off Guiana, 16.xi.1921, S. 200, 300 metres
of wire ; 1 specimen (now dry).

St. I223 iv , 22° 06' N., 84 58' W T ., off Cuba, i.ii.22, S. 200, 50 metres of wire ;
1 specimen (4-95 mm. in ventral mantle length).

New measurements have been made on the six larvae at my disposal.

No. 1. This specimen, in sea-water formalin, is now rather faded with faint
traces of chromatophores only on the third arms. The inierbrachial web is mode-
rately developed ; its depth in the A sector is about 3-0 mm., and about the same in
other sectors (except sector E, where it is distinctly less well developed).

No. 2 is a young and very interesting larva in which the third arm is more than twice
the length of the other arms ; the latter and the interbrachial web are but little
developed. These shorter arms have the thin whip-like tips so characteristic also

THE MACROTRITOPUS PROBLEM

Table IV.-

—Macrotritopus danae

Measurements in

mm.

No.

Ventral mantle length .

8-25

2'5

6*75

9-0

9-75

4-95

Dorsal mantle length .

io-8

3.0

n-55

13-35

I2-0

5-7

Mantle width

7'°5

2-4

6-o

7'35

7-8

3-6

Interocular width

6-45

2-55

4-8

6-75

6-o

4-2

Diameter of eye .

2*55

0*9

2-25

2-4

2-55

1*35

Arm length

ist arm

14-25

i'5

8-7

14-25

15-75

3-o

2nd arm

22-5

i-75

15-45

23-25

27*0

4-95

3rd arm

37*5

4'25

27-75

32-25

33-o

12-75

4th arm

18-0

i'75

11-25

21 -o

4-o5

Diameter largest sucker :

Diam. sucker, ist arm

•5-o -55

0-25

o-3

o-5

0-2

2nd arm

.55-0 -6

0-25

0-4

o-6

o-55

,, 3rd arm

.75-0-8

0-25

o-6

o-75

o-8

o-45

Indices :

Mantle arm index

22-0

58-8

24-35

27-9

29-6

38-8

Sucker index, 3rd

9-7

8-9

8-34

8-21

9-1

2nd .

7-28

5-93

6-67

5-64

—

ist .

6-67

4.44

5'5(>

5*13

4-04

of early stages of Octopus vulgaris, and their undeveloped state is reflected in the
number of suckers they carry. From the base of the arm there are 3-4 uniserial
suckers, 1-2 pairs of biserial suckers followed by rudiments of others. The third
arm, however, has about 7-8 well-formed, biserial suckers and rudiments of others
towards the tip. All the chromatophores have disappeared.

No. 3. In this specimen the right eye is torn and protruding (allowance has been
made for this in measurements) and the third left arm is mutilated. Except for
traces of two rows on the outer surface of the third arms the chromatophores have
disappeared.

No. 4. The mantle-shape in this specimen is as figured by Joubin & Robson. The
interbrachial web is moderately developed, the deepest sector being D. The chro-
matophore pattern has all but disappeared and there is no trace left on the mantle.
On the dorsal head there are two in the position indicated for M. scorpio by Berry
but the others are very faint. On the third arms there are traces of a few, opposite
suckers, and two faintly indicated rows on the outer surface of these arms.

No. 5. This paratype (B.M. 1929. 6. 29.1.) has retained far more of its colour
than any of the other "Dana" larvae ; this is perhaps partly due to the fact that the
specimen is in alcohol, and partly also because it has been kept for the greater part
of the time in total darkness. However, all chromatophores of the mantle have dis-
appeared except three in a row along the dorsal mantle edge and the deep-seated
ones of the visceral mass. Those of the dorsal side of the head can be counted and
conform to the vulgaris pattern.

The chromatophores of the third arms are contracted. On the outer surface there

7 6

THE MACROTRITOPUS PROBLEM

is, proximally, a single row, followed distally by a double row. Here and there there
is a suggestion of the pattern seen in M. kempt.

No. 6. No trace remains of the original chromatophore pattern.

(b) Macrotritopus kempi and other " Discovery " larvae

In addition to the syntypes of M. kempi, larvae have been found at the following
stations of the " Discovery " Expeditions (1925-1939) :

St. 276, 5 54' S., n° 19' E., N.W. of mouth of the Congo, 5.viii.i927, TYF,
150 (-0) m. ; 2 specimens of 3-3 and 4-5 mm. in ventral mantle length.

St. 276, N. 100 B, 110-0 m. ; 2 syntypes of M. kempi Robson and 1 specimen
of 3-75 mm. in ventral mantle length.

St. 290, 3 25' 25" N., 16 50' 52" W., off West Africa. 24. viii. 1927, N 100 B,
86-0 m. ; sounding 5,165 m. ; 1 specimen of 5-4 mm. in ventral mantle length.

St. 1592, 09 31' N., 17 37' W., off West Africa, 17.x. 1935, TYFB, 200-0 m. ;
1 specimen of 6-o mm. ventral mantle length.

St. 1594, 04 15' N., 12 58' W., off West Africa, 19.x. 1935, TYFB, 144-0 m. ;
1 specimen of 6-15 mm. in ventral mantle length.

St. 2646, 05 38' N., 14 03' W., off West Africa, 19.lv. 1939, TYFB, 250-0
m. ; 1 specimen of 3-2 mm. in ventral mantle length.

St. 2646, TYFB, 1,500-800 m. ; 1 specimen of 3-45 mm. in ventral mantle
length.
The standard measurements of the Discovery specimens are given in Table V.

Table V.

Measurements in mm.

No.

J 4

\
16

Ventral mantle length

8-i

7-65

3*3

4*5

3*75

5'4

6- 00

G-i 5

3-2

3-45

Dorsal mantle length

9-6

9'75

3'75

6-45

4'95

7-o5

6-75

9-0

3-9

4-2

Mantle width .

9'75

7-8

3-6

3-9

3-75

4-05

4-95

5-85

2-55

—

Interocular width

6-o

5*4

3'3

4'5

3'45

3 * 75

4' 6 5

5-7

2-70

3-°

Diameter of eye

1 -o

—

Arm length :

ist .

i3'5

11-25

2-25

4*2

2-25

5-4

5 ' 2 5

7-8

i-5

2-40

2nd .

22-2

18-0

3 '75

6-75

2-70

B-25

8 • 25

15-0

1-65

3-0

3rd .

32-25

—

10-95

16-5

10-5

16 -o

15-3

-^'5

6-o

6-75

4th .

18-45

14-25

2-25

4*5

2-85

6-5

8-25

10-05

i-5

2-25

Diameter, largest suckers :

2nd arm

o*45

o-35

0-2

0-25

0-2

o-3

0-4

—

3rd arm

o-55

o-5

o-35

0-4

o-35

o-45

o-55

o-3

Indices :

Mantle-arm index

25-15

—

30-I5

27-25

33-5

33-75

39-25

21-6

55 -o

51-0

Sucker index :

3rd arm

6-8

6-54

io-6

8-88

9-34

8-33

9-17

8-95

9-1

8-7

2nd arm

5-56

4'58

6-o6

5-55

5-33

3-42

5-o

0-5

6-82

—

Nos. 7 and 8 and the holotype and paratype of Robson's M. kempi, while Nos. 9, 10 and 11 are also from
the same station.

THE MACROTRITOPUS PROBLEM 77

The type is in very good condition, and so is the paratype except that in it both
third arms are damaged (the left one was regenerating a new tip). Robson's figure
(1929&, p. 170, fig. 60) gives a very good impression of the arrangements of chromato-
phores on the arms and head. On the ventral surface of the mantle, funnel and head
the pattern of chromatophores is very like that in 0. vulgaris. A notable feature
in this specimen is the development of the other arms (1,2 and 4), and the degree of
development of the conspicuous chromatophores on them as on the third arm.

On re-examination the head is seen to be rather contracted and the mantle fully
inflated in both specimens.

As regards measurements, in the species M. kempi the interocular width is very
much less than the mantle width — a feature of considerable importance to Robson
in defining his species. These specimens were picked out on board and transferred
to alcohol with consequent shrinkage of the head region ; this is evident from com-
parison with other specimens from the same station which were not picked out of
the plankton samples (PL 3, figs. 1 and 2). After making due allowances for size
(and refraining from comment at this point on the relative proportions of the arms)
all the specimens from this station undoubtedly form growth stages of one species.
The most constant and characteristic feature is the chromatophore pattern of the
third arms. On these (and on the second arms only to a lesser degree) the double
row of chromatophores on the outer surface is exceptionally clear and distinct,
especially the dorsal row. In this row each chromatophore is lengthened at right
angles to the axis of the arm, giving the arm the appearance of being banded. On
the sides of the third arm there is often a chromatophore at the base of each sucker
but sometimes this is placed between the suckers. On the ventral face of the third
arm there are also chromatophores between the bases of the suckers.

A re-examination of the original larvae lent by the Zoologiske Museum, K^ben-
havn, reveals little resemblance between the published figure (Degner, 1925, p. 79,
fig. 52) and the eight specimens available for study. In brief, the larvae are very
similar in general appearance to other Macrotritopus which have been described from
the Mediterranean and the Altantic (Text-figs. 1-5). Table VI gives the measure-
ments obtained in 1952.

Degner's specimen No. 2 (specimen H above) is the largest example in this series.
It has a ventral mantle length of 6-o mm. (given as 6-3 mm. by Degner ; the diffe-
rences between his measurements and mine are probably due to shrinkage in alcohol) .
Measurements of this and the other specimens from St. 184 are given in Table VI.

It will be seen (Text-fig. 5) that in this late larva and in the younger ones the shape
and proportions of the head and mantle are quite unlike Degner's fig. 52, and have
the usual form of Macrotritopus.

ZOOL. II, 4. 5§

THE MACROTRITOPUS PROBLEM

Overall length (i.e., including

3rd arm)
Ventral mantle length
Dorsal mantle length
Mantle width
Head width
Diameter of eye .
Arms :

1st

2nd

3rd

4th
Diameter proximal (basal)

sucker

Diameter last uniserial sucker

Diameter 1st biserial sucker .

Diameter largest sucker (3rd

arm) .

Table VI
Measurements in mm.

10-5
3'3
4-65
3-15
3'3
1 -05

i-8
2-4

6-75
i-8

0-2

0-25

o*3

10-95
3-9
4-85
2-85
2-85
i'3

2 -i

2-85

5-55
2 -i

0'2

0-25

4
5
3
3

85
o

55
9
3
5

2-55
3'i5
7-95
vo

10 -95

3'6

4-8

3'3
3-o

i'35

- /
5'i
2-55

>'22 5 —

)-25 —

)-2 5 —

0-3 0-3

8-55

7
9
7
55

o-8

0-2
0-2
0-2

8-7

2-85

4-o5

2-4

o-75

75
o

6 —

9
55
4
5

o
3

1 8

0-2
0-2

6-o

7^5

4*5
3'9
1-65

4-65
6-75

15-0
6-o

03
o-3

o-35

0-25 0-45

Specimens D and G are distorted, so that the mantle length for these is estimated.

The chromatophore pattern is rather faded but some features can be distinguished ;
those of the dorsal mantle have disappeared, but on the ventral surface there are
about 20, scattered over the surface as in 0. vulgaris larvae of similar size (i.e., in
mantle length). Deep-seated chromatophores are present on the dorsal head as in
the pattern typical of 0. vulgaris. On the arms the chromatophores occur in a single
row up to the edge of the web and then become double.

The web is fairly well developed and is of the order D = C>B>A = E. On
the first arm it reaches to almost half its length. The body and arms are covered
with the remains of Kollikersche buschel and some papillae, scattered over the mantle
are developing. There are 11 gill filaments in the outer demibranch of the left gill.

The feature of greatest interest is the third left arm, which, according to Degner,
was beginning to exhibit hectocotylization. I can make out no details in the struc-
ture of the tip of this arm, which is constricted off to form an egg-shaped structure
with a slightly-pointed tip ; its size is 0-30 mm. x 0-20 mm. (Text-fig. 6).

The identity of these larvae with Scaeurgus or Pter octopus depends on our inter-
pretation of this structure. Is it due to the vagaries of preservation, or can we inter-
pret its symmetrical shape and delicately rounded appearances as a developing
hectocotylus ? Consideration of this feature in relation to distribution is deferred
to p. 96.

(d) Specimens from Florida

Recently G. L. Voss (1951) has mentioned some Macrotritopus larvae which he
considered to be the scorpio form described by Berry (1920, p. 299). By courtesy

THE MACROTRITOPUS PROBLEM

Figs. 1-4. Scaeurgus sp. Degner, 1925 : the smaller larvae from "Thor" St. 184, Gulf of Corinth ;

all approximately x 6.
Fig. 5. Scaeurgus sp. Degner 1925: the largest larva from "Thor" St. 184 (specimen H), exhibiting

signs of hectocotylization ; x 6.
Fig. 6. Scaeurgus sp. Degner : tip of the third left arm from specimen H with rudimentary

hectocotylus ; greatly enlarged.

THE MACROTRITOPUS PROBLEM

"1 1 1 1 1 I I I

Macrotritopus

10
9

""• Degner (1925)

O Dana
"3 Discovery

<P Voss (1951)
(?) Berry (1920)

z
<

O
O

K9 cP°

p <t>o

VENTRAL MANTLE LENGTH

J I I I l_L

6 7 8 9 10

Fig. 7. Relationship between mantle width and ventral mantle length in Macrotri-
topus danae, M. kempi, M. scorpio, Scaeurgus sp. Degner and in additional specimens
from the "Discovery" Expeditions (1925-1929). The letter k denotes the type
material of M. kempi. All measurements in figs. 7-17 are in millimetres.

of Mr. Voss I have been able to examine two of the larval forms he mentions ; one
is in formalin and the other (No. 18) in alcohol.

I was unable to see much pattern on the surface of these larvae as their pigment
has largely disappeared, presumably through exposure to sunlight after fixation.
However, such pattern as remained indicated that there was nothing unusual about
it, and that it was probably no different in this respect from other Macrotritopus
larvae I had|seen.

THE MACROTRITOPUS PROBLEM

Table VII. — Macrotritopus larvae from Florida

Measurements in mm.

No.

Ventral mantle length

4'35

3-0

Dorsal mantle length

4.8

3-3

Mantle width ....

3*9

2-55

Interocular width

3'75

2-5

Arms :

ist

3-15

i'3

2nd .....

4'95

i*5

3rd

IO-2

6-o

4th

3-0

i*5

Diameter, basal sucker

0-2

0-125

second uniserial sucker

0-25

0-15

first biserial sucker

0-25

0-2

largest sucker, 3rd arm

I -2

0-3

d. A COMPARISON OF THE SPECIES

It was recognized quite early in the investigation that some or all of the Mediter-
ranean-Atlantic species of Macrotritopus might prove to belong to one or two species,
and the problem was to distinguish between growth stages and specific differences.
Accordingly some characters which might reveal significant points of difference were
examined.

Here it may be convenient to mention that ventral mantle length is used through-
out in all calculations involving mantle length. It appears more reliable than the
usually accepted dorsal mantle length (mid-point between the eyes to the apex of
the mantle), because retraction of the head in the larva can make as big differences
in measurements as distortion of the mantle margin! In the larva, too, there is a
distinct ridge on the dorsal mantle which provides a useful check if by chance the
ventral mantle edge is distorted.

All the accompanying graphs (except Text-figs. 9 and 10) are plotted on Wright-
man's double logarithmic paper to demonstrate the relationships of the various parts.
In using the data from this assorted collection of octopus larvae I have been very
conscious of the inadequacy of the material, and for this reason statistical methods
have not been exploited. It is, however, surprising in view of the assortment of
material how well grouped the measurements are in relation to straight lines.

It became obvious early in the investigation that material in alcohol, notably
Degner's Scaeurgus and Robson's two syntypes of Macrotritopus kempi, were much
shrunk, the shrinkage affecting different parts unequally. With age, too, shrinkage
progresses, and it must be remembered that the "Thor" collections were made 40-50
years ago and those of the " Discovery " between 1925 and 1939. Similarly the
material of M. danae and larvae from the " Discovery " collections, although both
are in formalin, are not strictly comparable from a preservation point of view, the
former being flaccid (and possibly a little swollen ?) by comparison with the latter,

THE MACROTRITOPUS PROBLEM

10
9
8

7
6

• Degner (1925)
O Dana
(J Discovery

QVoss (1951)
<*) Berry (1920)

t 1 r

Macrotritopus

1 1 I I I

o°
o o

96 o

°VS 3

VENTRAL MANTLE LENGTH

I 1 1)111

7 8 9 10

Fig. 8. Relationship between head width and ventral mantle length in Macrotritopus
danae, M. kempi M. scorpio, Scaeiirgus sp. Degner and in additional specimens from the
" Discovery " Expeditions (i 925-1 929).

which are in line condition. Widely different results can be obtained even with the
same medium from slight differences in the method of fixation, particularly in soft-
bodied animals without any substantial skeletal structures. The octopus arms in
particular are highly contractile, and it is by no means improbable that in their
extended condition they were two and even three times the length recorded after
preservation.

(a) Mantle length. At sizes below 4 mm. in ventral mantle length there is
very little difference between the length and the width of the sac. 1 Above this size

1 Mantle width, head width and sucker diameter grow in direct proportion to the length of the mantle
Text-figs. 7, 8 and 15).

THE MACROTRITOPUS PROBLEM 83

the mantle becomes distinctly elongated, except in the syntypes of M. kempi, where
it is decidedly inflated, giving an erroneous impression of mantle proportions (Text-
fig. 7, k). On the basis that all the specimens are growth stages, there are no signifi-
cant differences in the mantle proportions of the different species, although the
rather irregular plotting indicates that mantle width is not a useful diagnostic
character. This is to be expected, because the mantle sac is a highly muscular
pumping organ subject to alternate contraction and expansion.

(b) Head width. There are no features about head- width to suggest any signifi-
cant differences between the species (Text-fig. 8).

(c) The arms. The extraordinary length of the third arm in relation to arms
i and 2 is very striking in all the small and medium-sized larvae (Plate 3, figs. 3 and
4), but in the large ones the second arm may be 80% of the length of the third arm.
Earlier workers, notably Berry (1920) and Joubin & Robson (1929), were greatly
concerned with the relative proportions of these three pairs of arms, and although
they had some suspicions that discrepancies in relative arm lengths might be ac-
counted for by growth, nevertheless created species largely on these differences.

No one would deny that Degner's larvae belong to one species, although they
range from 2-7-6-0 mm. in ventral mantle length. By plotting the measurements
of the arms in relation to ventral mantle length in a simple graph we get the results
shown in Text-fig. 9 ; this clearly indicates that differences in arm lengths may be
accounted for by growth. A similar impression is gained from plotting the measure-
ments for " Dana " and " Discovery " specimens in the same way (Text-fig. 10).

In the youngest known stages (2-5-3-0 mm. in ventral mantle length) the arms
are already differentiated in length with the third arms greatly exceeding the second
pair, which in turn are a little longer than the first pair. Between 3-0 and 4-0 mm.
the difference in length between the first and second arms becomes more marked,
and beyond this all three arms (1, 2 and 3) are growing rapidly.

These simple graphs suggest that the changes in relative proportions of the arms
may indicate stages in allometric growth. In Text-figs, n-13 the arm-lengths
(arms 1, 2 and 3) are plotted according to the species or source of the material and
for each we get a reasonable plotting, which again suggests that they all belong to
one species, and it can be said that in all three arms growth is proportional to the
mantle growth.

When the results for the three pairs of arms are grouped, the resulting graph
demonstrates that all three grow at approximately the same rate (Text-fig. 14). The
rate of growth is indicated by the slope of the mean, and it will be seen that the growth
rate is the same in arms 1 and 3, but arm 2 maybe growing at a slightly faster rate
than the other two. The differences in arm length are dependent on when growth
begins ; in arm 3 it begins at a ventral mantle length of about 1-4 mm., that is, at
hatching size if we assume that the larva comes from a typical octopus egg. The
second arm does not begin to grow until the larva has a ventral mantle length of
about 2-3 mm. ; it is closely followed by the first arm when the ventral mantle
length reaches 2-5 mm.

8 4

THE MACROTRITOPUS PROBLEM

I
H

-J

12 k

lOh

1 1 1 1

• • Degner (1925) 4

-A Berry (1920)

m O Voss(l95l)

—

^■™

— <

—

•

- p

• f

—

•

—

•

1 °

ii 6 ▲

1 1

8 -

7 •

6 ~

VENTRAL MANTLE LENGTH

Fig. 9. Length of arms 1, 2 and 3 in relation to ventral mantle length in Scaeurgus sp.
Degner (filled circles) and Macrotriiopus scorpio Berry (open circles and triangles),
plotted on ordinary graph paper.

THE MACROTRITOPUS PROBLEM

1 1

Macrotritopus

30 ~

O Dana

Discovery

25- cs

20 k ^

X
h-
KD

15 \- -J

■

10 -

O H

• 4

2 3 5 7 9

VENTRAL MANTLE LENGTH

Fig. 10. Length of arms 1, 2 and 3 in relation to ventral mantle length in material from
the Danish " Dana " Expeditions 1921-1922 (open circles) and from the " Discovery "
Expeditions, 1 925-1 939 (filled circles), plotted on ordinary graph paper.

THE MACROTRITOPUS PROBLEM

1 1

1 1 1 1 1 1

Macrotrit

opus

•

Degner (1925)

Dana

8°

Discovery

10
9

Voss (1951)

Berry (1920)

—

»<1

•

—

*->

— u_

•••

*••

_ z

-J

• • •

•

O 9

. .1 . 1

. 1 1 1 1 1 1

2 3 4 56789 10

VENTRAL MANTLE LENGTH

Fig. ii. Relationship of the length of the first pair of arms to ventral mantle length in
Macrotritopus danae, M. kempt, M. scorpio, Scaeurgus sp. Degner, and in additional
specimens from the " Discovery " Expeditions (i 925-1 939).

(d) Sucker diameter. The diameter of suckers of the third arm have been
plotted in relation to ventral mantle length (Text-fig. 15) ; this graph shows more
clearly than the others the shrinkage caused by storage in alcohol compared with
formalin preservation. Most of the Scaeurgus larvae and the syntypes of Macro-
tritopus kempt (marked k) fall well below the line. Apart from this the graph reveals
no significant differences between the species.

THE MACROTRITOPUS PROBLEM

• Degner (1925)
O Dana

1 1

Macrotrltop

1 l 1 l 1 1

L'S

—

O Discovery
(j) Voss (1951)
(f) Berry (1920)

3
30

—

9
8
7

■■■

~ mm

6 3

"" H

•

— <

-0

c
— r^

—

^ •

— LU

—

—i

•-•

O •

VENTRAL MANTLE

LENGTH

1 1

1 1 1 1 1 1

2 3 456709 10

Fig .12. Relationship of the second pair of arms to ventral mantle length in Macrotvilopus
danae, M. kempi, M. scorpio,. Scaeurgus sp. Degner, and in additional material obtained
by the " Discovery " Expeditions (1925-1939).

(e) The radula. The radula of a specimen from " Discovery " St. 1592 has been
examined. It exhibits a symmetrical A 3~4 seriation of the cusps of the rhachidian,
but there is some irregularity (perhaps to be expected in a young specimen). The
first lateral tooth is narrow with a short pointed cusp, and the second lateral has a

THE MACROTRITOPUS PROBLEM

50
40

30
20

Macrotrhopus

O Dana

3 Discovery

-Q Voss (1951)

(J) Berry (1920)

3QO

«o

»3

o
9

VENTRAL MANTLE LENGTH
I I l I I I I I

2 3 4 5 6 7 8 910

Fig. 13. Relationship of the third pair of arms to ventral mantle length in Macrotri-
topus danae, M. kempt, M. scorpio, Scaeurgus sp. Degner, and in additional material
obtained by the " Discovery " Expeditions (1 925-1 939).

THE MACROTRITOPUS PROBLEM

|—

— "1 " ! 1

Macrotritopus

1 1

1 1 1

—

40
30

" O
" ©

1st i
2nd

a^rm

ARM

••

•
• ••

—

_ • 3rd.

ARM

•

8°

—

•••

9
8
7
6

•

u
r

•
O

«• ® ®

"• # ®®°g

••

—

3
2

— _l

°
®®

®o
cP

1 1 1

1- L

1 1 1

5 6 7 8 9 10

VENTRAL MANTLE LENGTH

Fig. 14. Relationship of arms 1, 2 and 3 in relation to ventral mantle length in all the specimens

examined.

THE MACROTRITOPUS PROBLEM

i_
m

4 -

T 1 — 1 I I I

Macrotritopus

# Degner (1925)

3 Discovery
<§> Voss (1951)

oo
o

m> K

3 3

VENTRAL MANTLE LENGTH

1— JL-I I I-

6 7 8 9 10

Fig. 15. Relationship of the sucker diameter of the third arm to ventral mantle length in
Macrotritopus danae, M. kempi, M. scorpio, Scaeurgus sp. Degner, and in additional
material obtained by the " Discovery " Expeditions (1925-1939). The letter k
denotes the type material of M. kempi.

well formed endocone but no ectocone. The third lateral is distinctly long, slender
and rather curved. The marginal plates are lozenge-shaped, a little broader than
long.

The radula of the holotype of Macrotritopus kempi has also been examined and is
very similar to the above, although details of the sedation of the cusps are difficult

THE MACROTRITOPUS PROBLEM 91

to interpret. Even though the preparation is a poor one, the cusps have a symme-
trical arrangement and not an asymmetrical one as stated by Robson (1929, p. 312).
The laterals are very similar to those of the specimen from " Discovery " St. 1592.

(f) Gill filaments. The number of gill filaments per demibranch has been
noted in one specimen of each of the following : Scaeurgus sp. Degner, E. Mediter-
ranean, n filaments ; Macrotritopus kempi Robson, West Africa, 11-12 filaments ;
Macrotritopus danae Joubin & Robson, W. Mediterranean, at least 11 ; Macrotri-
topus danae Joubin & Robson, off Cuba, at least II.

The third and fourth specimens listed above have deteriorated considerably in
formalin, and the detailed structure of the gills, is not now very satisfactory for
counting filaments.

All the evidence suggests that all the Macrotritopus larvae examined belong to a
single wide-ranging species. I have not seen VerruTs Macrotritopus (M. equivocus
Robson), but there appears to be no doubt about its identity with the other larvae ;
it appears to be nothing more than a very large larva carried by water movements
well beyond its normal range.

e. THE BREEDING SEASON

On the assumption that all the larvae belong to a single species it is possible to get
some indication of the breeding period. We have to assume that once the eggs are
laid (presumably at moderate depth) they take at least as long, if not longer, to
hatch as in a littoral octopod. In Octopus vulgaris this is a period of 21-30 days,
according to temperature, and it may be even longer in a deep-water species where
sea temperature on its spawning grounds is likely to be lower.

In order to supplement the data obtained from the material studied, I have
drawn on a list of occurrences given by Joubin (1937, p. 36) in addition to my own
material. In the latter (Text fig. 16) ventral mantle length is taken as a criterion of
size, while in the former overall length appears to be quoted (Text-fig. 17) . As breeding
may be largely controlled by temperature, the Mediterranean records have to be
treated separately from those of the tropical Atlantic.

(a) The Mediterranean. In this warm temperate area we have records of
larvae of all sizes but only for the months of August and September. It is true that
during the Third " Dana " Expedition in 1921-22 the Mediterranean was visited
only between 21st September and 4th October, 192 1. The earlier expedition in the
' Thor " (1908-1910) did, however, cover a much longer interval of time (14th
December, 1908, to 21st February, 1909, and 26th June, 1910, to 7th September,
1910) but recorded larvae only in August.

If, as the meagre evidence suggests, larvae are common only in August and Sep-
tember, spawning must take place in June-July and possibly in August as well. By
the end of September, when the largest larvae have been found, the majority are
ready for settlement on the bottom, while a few late larvae may persist well into
October. This is in agreement with Naef 's statement that octopod larvae at Naples

THE MACROTRITOPUS PROBLEM

— i t r

— O Atlantic

— # Mediterranean

I '

LU 8

—J

z
<

£ 5

> 4h
3
2 -

t — i r

J L

JAN. FEB. MAR. APR. MAY JUNE JULY AUG. SEPT. OCT. NOV. DEC.

Fig. i 6. Size (ventral mantle length) in relation to seasonal occurrence in the Atlantic and

Mediterranean.

are more common in August and September than during other times of the year
and suggests summer breeding.

(b) The Atlantic. In the Atlantic only two larvae have been found to the
north of 30 ° N. They were found at the following positions :

"Dana" St. 1152, 30 17' N., 20 44' W., S.W. of Madeira, 23.X.1921;
1 specimen of 18 mm.

" Dana" St. 1341, 33 15' N., 68° 20' W., between Bermuda and the mainland,
14. v. 1922; 1 specimen of 20 mm.

All the remainder were found in tropical or sub-tropical waters off West Africa
and on the American coast, mainly off islands of the Greater and Lesser Antilles.

On the African coast the ships of the " Discovery " Committee took specimens in
March, April, August and October during transit voyages to and from the Antarctic.
There has, however, been no continuous sampling on this coast to give an accurate
picture of seasonal distribution.

THE MACROTRITOPUS PROBLEM

In the Central American region however the " Dana " was constantly taking
plankton hauls between November, 1921, and May, 1922. There was no month
in this period when larvae were not present. This implies that the breeding period
was spread over the whole of this time, and if we take the records from West Africa
and Florida into consideration, June is the only month for which larvae have not
been recorded, and this apparent absence seems to be due to lack of observations.

.... ( .

— •
•

1 1

— •
•

O CO

—

jE 40
O

ULI

_l
30

- ••

)

•

—

- •

„ 1

1 1

SEPT. OCT. NOV. DEC. JAN. FEB. MAR. APL. MAY JUNE

Fig. 17. Size (overall length) in relation to seasonal occurrence in the Atlantic and
Mediterranean based on records of Macrotritopus spp. from the cruises of the " Dana
(1921-1922).

It is therefore clear that breeding may be noted in tropical and sub- tropical waters
throughout the year. The greatest number of larvae (31) was taken in December,
but the significance of this is not known. It may be that plankton hauls during this
month were made in an area where there was concentration of breeding adults, or
it might mean that this was the period of maximum breeding activity. If the latter it
would imply maximum breeding in winter in the tropics — contrasting with summer
breeding in the more temperate Mediterranean. In the Tropics, however, the factors
which may exercise a big influence on temperature (which governs breeding) are
local hydrographic conditions. For an animal which lives on the narrow coastal
shelf (as off West Africa and the Antilles) local temperature conditions are bound to
vary enormously according to movements of water masses, especially upwelling.
Local breeding may therefore vary according to place and season.

94 THE MACROTRITOPUS PROBLEM

r. THE SIGNIFICANCE OF THE M acrotritopus LARVA

The plankton life of the larvae of Octopus vulgaris (to judge from results hitherto
unpublished) may have a duration of at least four weeks, and possibly eight or ten
weeks under normal hydrographic conditions. For M acrotritopus it is reasonable
to suppose a planktonic life of longer duration (perhaps for three to five months) on
the basis that the larvae reach a much larger size. There is, however, another
factor to be considered ; one of Degner's larvae from the comparatively shallow
Gulf of Corinth was developing a hectocotylus at 6-o mm. in ventral mantle length.
To judge from evidence from all sources, including Naef's work in the Mediterranean,
metamorphosis (here interpreted as the abandonment of a planktonic or pelagic
phase) also means the beginning of sex differentiation and the development of
gonads ; in the male the visible outward sign is the development of a hectocotylus.
In the Gulf of Corinth the water is not very deep, and a larva ready to become ben-
thic should find a suitable bottom at the right time, so that there would be no great
delay in settlement. This implies that the ordinary M acrotritopus larva completes
its planktonic phase at 6-7 mm. in ventral mantle length (allowance being made for
the shrinkage of this material).

Other Macrotritopus larvae, however, reach a much bigger size without developing
any sign of hectocotylus or of gonads. In M. equivocus the ventral mantle length
is estimated to be n mm. ; in M. kempi and M. danae respectively it is 8- 1 mm. and
9-75 mm. In view of the fact that I can find no other differences between these
larvae and Degner's species, some other explanation must be sought for the non-
appearance of signs of sexual differentiation or maturation, and I am inclined to
interpret these large larvae as extended pelagic phases.

This phenomenon has been noticed before in marine invertebrates, but its signifi-
cance does not seem to have been discussed in relation to distribution and dispersal.
It is known from the work of Dr. Marie V. Lebour that in many molluscs a creeping-
swimming stage of several days' duration is reached during metamorphosis, which,
as suggested b}/ Wilson (1937), may enable the larva to persist in a state whereby it
can be swept away by currents to a more suitable substratum. From our present
knowledge of the change to a benthic stage it is evident that species with a quick
metamorphosis can delay settlement for several days, thus giving them a better
chance of rinding a suitable bottom.

When metamorphosis is gradual, settlement can be delayed for a long period, as
has been demonstrated by Day & Wilson (1934) for the polychaete Scolecolepsis
fulginosa. During this period of uncertainty the worm continues to grow beyond
the 14-setiger stage at which these authors found metamorphosis to take place. In
another polychaete, Loimia medusa, Wilson (1928) found that its planktonic exist-
ence ended when the larva had reached a length of about 6-o mm. At this size the
first pair of gills is beginning to divide, the second is only a rudiment and the third has
not made its appearance. Still larger larvae were referred to the species by Monro
(1930 and 1 931) ; of these the largest example from " Discovery " St. 102 was 15
mm. long and had three pairs of well-developed gills. If we accept the view (as is

THE MACROTRITOPUS PROBLEM 95

most likely) that these larvae belong to the widely-distributed Loimia medusa, they
can only be interpreted as an extended pelagic phase.*

Gurney (1942, pp. 71-75) gives details of several larvae of decapod Crustacea which
have been caught at sizes well above the known size for moulting into post-larvae.
He was " reluctant to accept any theory of abnormality " for these giant larvae, and
supposes that they had failed to metamorphose at the right time and continued to
grow." He adds : ' ... in a situation like that of Bermuda, where abyssal
depths are so close to the shore, it must be quite usual for larvae to find themselves
over deep water at the time of metamorphosis."

Similarly when the distribution of Macrotritopus larvae is considered many of
them were also taken fairly close to land, but in deep water off the edge of the con-
tinental shelf. At the places where they were taken (off West Africa and in the
Antilles) the continental shelf is very narrow, so that the chances of larvae being
swept over deep water are great. On the West coast of Africa, particularly, this
seems to occur, and Knudsen (1950) attributes the high percentage of gastropods
with very short planktonic or non-pelagic development on this coast to the influence
of water movements in eliminating species with planktonic stages of long duration.

The Macrotritopus larva with its long third arms (which may be as useful as the
squid's tentacles in catching food), its narrow squid-like profile and its curious colour
pattern recalling that of the epipelagic larvae of the Cranchiidae, seems well adapted
for a pelagic life. It may be that Macrotritopus specimens of above 6-7 mm. in
ventral mantle length are larvae which have grown large in the plankton because
they have not been able to find a suitable bottom (or a suitable bottom at the right
depth) for metamorphosis. In other words delayed settlement has resulted in
an extended pelagic phase which may be of considerable significance in dispersal.
What is important is that it can travel immense distances in the plankton. The
M. equivocus found off Cape Sable, Nova Scotia, had reached a mantle-sac length of
n mm., and must have been transported for hundreds of miles by currents (probably
from the Carribean or Florida) ; it was likely to be well out of the area in which it
could survive as an adult. It may, however, be pertinent to draw attention to the
role of the extended pelagic phase in maintaining uniformity in a wide-ranging species.

The remarkable similarity of populations of Scaeurgus from the Atlantic and
Pacific was noted by Berry (1914, p. 306), who adds (in litt., quoted by Voss, 1951) :
"if it really is all one species, then there must be some special reason for lack of
divergent speciation, and I can conceive no sensible explanation for this in a bottom
form except for the possession of a planktonic larval stage of some duration."

g. THE IDENTITY OF Macrotritopus WITH Scaeurgus

When it became obvious that all the Altlantic Macrotritopus were referable to a
single species, it was deemed possible to suggest a possible adult with which they

* Dr. Gunnar Thorson has kindly drawn my attention to Lemche's observations on the North
Atlantic tectibranch Diaphana minuta Brown, which seem to suggest that the larva of this mollusc
is able to delay metamorphosis when swept our over deep water (Thorson, 1946, p. 466 and Lemche,
1948, p. 9). In shallow water adults the embryonic shells are much smaller than those in specimens
from depths exceeding 1,000 m. These larger embryonic shells are believed to indicate delayed settle-
ment and resulting growth of the larval shell in specimens which have metamorphosed in deep water.

96 THE MACROTRITOPUS PROBLEM

could be linked. When this paper was in early stages of preparation in 1950 there
were four species with one of which they could possibly be linked, namely, Octopus
vulgaris, 0. macropus, Scaeurgus unicirrhus and Pteroctopus tetracirrhus.

The first, 0. vulgaris, was the only species which had a known distribution corre-
sponding most closely to that of the larvae (Macrotritopus equivocus being con-
sidered to have been carried well out of its normal distribution area). The details
of the structure of the radula — asymmetrical in 0. vulgarus and symmetrical in
Macrotritopus — together with the occurrence of a normal larva of vulgaris in the
same hauls as M. kempi (St. 276) were sufficient evidence that 0. vulgaris could not
be the adult.

The second, 0. macropus Risso, has a similar but a more restricted distribution in
the Mediterranean and both sides of the tropical Atlantic. This species, too, was
ruled out because the details of the radula showed specific differences, and also
because larvae of 0. macropus were positively identified during the investigations.

Scaeurgus unicirrhus and Pteroctopus tetracirrhus were originally included in the
list of possible adults, not because of any great similarity between the known dis-
tribution of the adults and that of Macrotritopus, but solely because each species
had a sinistral hectocotylus. Degner's largest larva (specimen H) had a swelling
at the tip of the third left arm, and on re-examination of the species I found it
difficult to imagine its delicately rounded structure being the result of any mal-
formation of the tip of the tentacle, and had to conclude that it was a true rudiment
of a hectocotylus.

Evidence from distribution of the adults was not encouraging in 1950, especially
as neither species had been found on the American side of the Atlantic and records
outside the Mediterranean were very few. However, Voss (195 1) gave the first
records of Scaeurgus from the Western Atlantic (Florida), and Adam (1952) recorded
Pteroctopus from West Africa. Full records of the two species outside the Mediter-
ranean (where both are found) are given below :

Scaeurgus unicirrhus : Pacific Ocean ; Hawaiian Islands (Berry, 1913 and
1914 as patagiatus) ; South of Kyushu (Sasaki, 1920 and 1929 as patagiatus) ;
Indian Ocean ; Saya de Malha (Robson, 1921) ; Atlantic Ocean, Florida (Voss,

1951).

Pteroctopus tetracirrhus : Atlantic Ocean ; Cape Verde Islands (Fischer &

Joubin, 1906 and 1907) ; Azores (Joubin, 1900) ; off the coast of Africa (Adam,

1952).

Scaeurgus appears from our present knowledge of distribution to have a much

wider range than Pteroctopus ; as both species live on moderately deep bottoms, it

it not surprising that there are so few tropical records in view of the difficulties of

operating trawls in areas where the bottom is covered with living or dead coral.

Both these species have a symmetrical arrangement of cusps on the rhachidian
tooth of the radula as in Macrotritopus. A closer examination of the Macrotritopus
radula shows that it agrees well with Scaeurgus, and differs from Pteroctopus in having
a narrower first lateral tooth, a longer, more slender and more curved third lateral,
and lozenge-shaped marginal plates. In Pteroctopus (see Adam, 1952, fig. 55) the
marginal plates are wider and slightly curved in outline.

THE MACROTRITOPUS PROBLEM 97

All the evidence suggests that Macrotritopus is the larval stage of Scaeurgus
unicirrhus, and at present there is no indication that more than one species is in-
volved. Scaeurgus has been found in the warm waters of all oceans, and its wide
distribution (surprising in a bottom-dwelling animal) has clearly been achieved by
this striking pelagic larva. The number of gill filaments per demibranch in Scaeurgus
(11-14) is exceptionally high for a benthic octopod living at considerable depths
near the edge of the continental shelf, but is probably related to the pelagic require-
ments of the larva.

h. CONCLUSION

All four Atlantic species of Macrotritopus have been demonstrated to belong to
one species, also now linked with its adult, Scaeurgus unicirrhus (Delle Chiaje).
Macrotritopus falls into the synonymy of Scaeurgus, and the position regarding the
species is summarized as follows :

Scaeurgus unicirrhus (Delle Chiaje).

Octopus gracilis Verrill, 1884, p. 236 non Eydoux & Souleyet, 1852, p. 13.

Macrotritopus gracilis, Grimpe, 1922.

Macrotritopus equivocus Robson, 19290, p. 311.

Octopus (Macrotritopus) equivocus, Robson, 1929&, p. 169.

Octopus (Macrotritopus) equivoca, Joubin & Robson, 1929, p. 93 ; Joubin,

*937> P. 34-
Polypus scorpio Berry, 1920, p. 299.
Octopus (Macrotritopus) scorpio, Robson, 19296, p. 169 ; Joubin & Robson,

1929, p. 93.
Macrotritopus kempi Robson, 19290, p. 311.
Octopus (Macrotritopus) kempi, Robson, 1929&, p. 170.
Octopus (Macrotritopus) danae Joubin & Robson, 1929, p. 8y ; Joubin, 1937,

P- 33-
Macrotritopus spp. Joubin, 1937, p. 36.
Scaeurgus (Troschel sp. juv.) Degner, 1925, p. 79.

The great length of the third arm is a larval character only, for it is not found
in the adult, in which the arms are subequal. The long third arm may have some
special significance in larval life, serving perhaps for capturing food in the same way
as the tentacles of the epipelagic squids. The high number of gill filaments, too,
indicates an active larval existence which is in contrast with the low number found
in octopods like Bathypolypus and Bentheledone, which are benthic deep-water
octopods of the continental slope.

The records indicate that the Macrotritopus larva lives in the upper 200 metres.
Most of the hauls from deeper water were made with open nets, so that positive
conclusions cannot be reached from them, but the specimen taken by the " Dis-
covery " with a closing net between 1,500 and 800 metres was at a surprising depth
in a species which has all the appearance of being a surface form. Its occurrence
at this depth suggests that it was seeking bottom preparatory to metamorphosis.

If we accept the occurrences of Macrotritopus larvae in the vicinity of land (or

98 THE MACROTRITOPUS PROBLEM

near the edge of the continental slope) as evidence of the existence of benthic adults
nearby, Scaeurgus has a more general distribution in the warmer waters of the
Atlantic than hitherto supposed. As already mentioned, the difficulties of trawling
over coral in the tropics probably accounts for the scarcity of records of the adult.

REFERENCES

Adam, W. 1952. Cephalopodes. Res. Sci. Exped. Oceanog. Beige Eaux Cotieres Africaines
de V Atlantique Sud (1948-1949), 3 (3) : 1-142, pis. 1-3 and text-figs. 1-57.

Berry, S. S. 1913. Some new Hawaiian Cephalopods. Proc. U.S. Nat. Mus. 45 : 563-566.

1914. The Cephalopoda of the Hawaiian Islands. Bull. Bur. Fish. Washington, 32

(1912): 257-362, 40 text-figs, and 11 pis.

1920. Preliminary diagnoses of new cephalopods from the Western Atlantic. Proc. U.S.

Nat. Mus. 58 : 293-300.

Day, J. H., & Wilson, D. P. 1934. O n tne relation of the substratum to the metamorphosis

of Scolecolepis fuliginosa (Claparede). /. Mar. Biol. Assoc. 19 : 655-662.
Degner, E. 1925. Cephalopoda. Danish Oceanog. Exped., 1908-1910, 2, C. 1 : 1-94, 52

text-figs.
Fischer, H., & Joubin, L. 1906. Note sur les Cephalopodes captures au cours des expedi-
tions du Travailleur et du Talisman. Bull. Mus. Hist. Nat. 1906, No. 4 : 202-205.

1907. Cephalopodes. Exped. Sci. Travailleur et du Talisman, 8 : 313-353.

Grimpe, G. 1922. Systematiche Ubersicht der Europaischen Cephalopoden. Sitzber. Naturf.

Ges. Leipzig, 9 : p. 36.
Gurney, R. 1942. Larvae of decapod Crustacea. Ray Soc. 1942, vi -f 306 pp., 122 text-figs.
Joubin, L. 1900. Cephalopodes provenant des Campagnes de la Princess Alice. Res. Camp.

Sci., Monaco, 17 : 1-135, 15 pis.
— ■ — ■ 1937. Les octopodes de la croisiere du ." Dana " 1921-22. Dana Rep. No. 11 : 1-49,

53 text-figs.
— ■ — ■ & Robson, G. C. 1929. On a new species of Macrotritopus obtained by Dr. J. Schmidt's

" Dana " Expedition, with remarks on the genus. Proc. Zool. Soc. London, Pt. 1 : 89-94,

1 text-fig.
Knudsen, J. 1950. Egg capsules and development of some marine prosobranchs from

tropical West Africa. "Atlantide" Rep. No. 1 : 85-130, 31 text-figs.
Lemche, H. 1948. Northern and Arctic tectibranch gastropods. I. The larval shells.

II. A revision of the cephalaspid species. A'. Danske Vidensk, Selsk. Biol. Skr. 5 (3) :

1-36, 80 text-figs.
Monro, C. C. A., 1931. Note on the pelagic phase of a polychaete worm belonging to the

family Terebellidae. Ann. Mag. Nat. Hist. (10), 7 : 212-215.

■ 1936. Polychaete Worms, II. "Discovery " Rep. 12 : 59-198, 34 text-figs.

Naef, A. 1921-28. Die Cephalopoden. Faune et Flora del Golf di Napoli. Monographic, 35.
Pickford, G. E. 1945. Le poulpe americain : a study of the littoral Octopoda of the Western

Atlantic. Trans. Conn. Acad. Arts Sci. 36 : 701-81 1, 14 pis.

1947- A review of the littoral Octopoda from Central and Western Atlantic Stations in

the collections of the British Museum. Ann. Mag. Nat. Hist. (11) 13 : 412-428, 1 text-fig.

Rees, W. J. 1950. The distribution of Octopus vulgaris Lamarck in British Waters. /. Mar.
Biol. Assoc. 29 : 361-378, 3 text-figs, and 3 pis.

1952. Octopuses in the Channel. New Biology, No. 12 : 58-67, 3 text- figs.

Robson, G. C. 1 92 1. On the Cephalopoda obtained by the Percy Sladen Trust Expedition

to the Indian Ocean in 1905. Trans. Linn. Soc. (2) Zool. 17, (4) : 429-442, 2 pis.

1929a. Notes on the Cephalopoda : VII. On Macrotritopus Grimpe with a description

of a new species. Ann. Mag. Nat. Hist. (10) 3 : 311-313.

■ — — 1 929ft. A Monograph of the Recent Cephalopoda. Parti: Octopodinae. Brit. Mus. (Nat.
Hist.), London, xi + 236 pp., 89 text-figs, and pis. 1-7.

THE MACROTRITOPUS PROBLEM 99

Sasaki, M. 1920. Report on Cephalopods collected during 1906 by the United States Bureau

of Fisheries steamer " Albatross " in the North-western Pacific. Proc. U.S. Nat. Mus.

57 : 163-203, pis. 23-26.
1929. A monograph of the dibranchiate cephalopods of Japanese and adjacent waters.

/. Coll. Agri. Hokkaido Univ. 20, Supplementary Number, 1-357, 3° pis. an d 159 text-figs.
Schmidt, J. 1929. Introduction to the Oceanographical Reports, including list of the stations

and hydrographical observations. Danish " Dana " Expeditions 1920-22, No. 1 : 1-87,

18 text-figs, and 6 pis.
Thorson, G. 1946. Reproduction and larval development of Danish marine bottom inverte-
brates, with special reference to the planktonic larvae in the Sound (^resund). Medd.

Komm. Danmarks Fisk. Havunderseg. Ser. Plankton. 4 (1) : 1-523, 199 text-figures.
Verrill, A. E. 1884. Second Catalogue of Mollusca recently added to the fauna of the New

England Coast and the adjacent parts of the Atlantic, consisting mostly of Deep-sea

species, with notes on others previously recorded. Trans. Conn. Acad. Sci. 6 : 139-294,

pis. 28-32.
Voss, G. L. 195 1. A first record of the cephalopod, Scaeurgus unicirrhus, from the Western

Atlantic. Bull. Mar. Sci. Gulf Caribbean, Coral Gables I (I) : 64-71, 2 text-figs.
Wilson, D. P. 1928. The post-larval development of Loimia medusa Sav. /. Mar. Biol.

Assoc. 15 : 129-146, 2 pis. and 7 text-figs.
1937. The influence of the substratum on the metamorphosis of Notomastus larvae.

Ibid. 22 : 227-243.
— ■ — " 1952. The influence of the nature of the substratum on the metamorphosis of the larvae

of marine animals, especially the larvae of Ophelia bicornis Savigny. Ann. Inst. Oceanog.

Paris, 27 (2) : 49-156, 2 text-figs.

PRESENTED

r6?EP 1954

PLATE 3.

Figs, i and 2. Macrotritopus larva of 3-45 mm. in ventral mantle length from "Discovery"
St. 2646.

Figs. 3 and 4. Macrotritopus larva of 375 mm. in ventral mantle length from " Discovery "
St. 276. This specimen was taken in the same haul as Robson's syntypes of M. kempi.

Bull. B.M. (N.H.) Zoology II, 4.

PLATE 3

'■■... "' ■-

■

*'/M'\ .

' '•" *f '/

>i , s

^P *wf

jf \/

P / ]

■

MACROTRITOPUS LARVAE.

PRESENTED

3 FEB 1954

\ Hv&miZk

W&3

PRINTED IN GREAT BRITAIN BY
ADLARD AND SON, LIMITED,
BARTHOLOMEW PRESS, DORKING

HEARING IN CETACEANS

PRESENTED

2 8 JAN 1954

F. C. FRASER

AND

P. E. PURVES

BUELET1N OF
THE BRITISH MUSEUM (NATURAE HISTORY)
ZOOLOGY Vol. 2 No. ^

LONDON: 1954

HEARING IN CETACEANS

F. C. FRASER

AND

P. E. PURVES

Pp. 101-114 ; Pis. 4-5

BULLETIN OF
THE BRITISH MUSEUM (NATURAL HISTORY)
ZOOLOGY Vol. 2 No. S

LONDON: 1954

THE BULLETIN OF THE BRITISH MUSEUM
(NATURAL HISTORY), instituted in 1949, is
issued in five series corresponding to the Departments
of the Museum, and an Historical Series.

Parts appear at irregidar intervals as they become
ready. Volumes will contain about three or four
hundred pages, and will not necessarily be completed
within one calendar year.

This paper is Vol. 2, No. 5 of the Zoological
series.

PRINTED BY ORDER OF THE TRUSTEES OF
THE BRITISH MUSEUM

Issued January, 1954. Price Four Shillings.

HEARING IN CETACEANS

By F. C. FRASER and P. E. PURVES.

SYNOPSIS

Essential anatomical features of the region of the ear of cetaceans are described and figured.

Earlier hypotheses relating to the mode of hearing in cetaceans are briefly reviewed and
criticized.

Reasons are given for regarding the method of hearing in cetaceans as being essentially
similar to that of terrestrial mammals with certain anatomical modifications required for
receiving water-borne vibrations.

A simple experiment is described demonstrating the superiority of the walls of the meatus
over other parts of the body as a conductor of longitudinal vibrations.

At the present time the sounds emitted by cetaceans and consequently their sense
of hearing are the subject of a considerable amount of investigation. The remarks
that follow are in anticipation of a more detailed account of the cetacean ear which
the writers have in preparation.

In order to clarify the hypothesis arrived at a brief summary of the anatomical
features involved is necessary (in connection with which reference to Plates i and 2
should be made).

The external auditory meatus is a continuous narrow tube in the toothed cetaceans
(with the exception of Physeter catodon, see Clark, 1948). In the whalebone whales
it is usually closed for some part of its length immediately internal to the blubber
layer. The meatus is lined by a pigmented extension of the epidermis. Surrounding
the lining layer is a fibro-elastic sheath, the fibres of which run predominantly along
the length of the meatus. Surrounding this sheath again is a fibro-cellular structure
in which a thin stratum of circular constrictor muscle has been observed.

Associated with the tube are incompletely investing cartilages into which auricular
muscles are inserted. Of these one of the most important is the m. auricularis-
occipitalis-profundus which Bonninghaus (1904) described and figured in detail in
connection with the dissection of a Common Porpoise. Bonninghaus' figure shows
the acute dorsally- and slightly caudally-directed bend into which the meatus is
thrown in its course. The muscle just referred to originates at the top of the skull
and extends downwards to its insertion at the apex of the bend in the meatus.
Another equally important muscle is Bonninghaus' m. zygomatico-auricularis
which has its origin on the zygomatic process of the squamosal and converges to be
inserted into the cartilage at the inner end of the meatus.

In both toothed and whalebone whales the meatus widens out mesially to ter-
minate at the tympanic membrane. The latter is not as in terrestrial mammals
a thin translucent membrane. It consists of two parts, a fibrous region containing

ZOOL. 2, 5.

io 4 HEARING IN CETACEANS

radial and concentric fibres and a fibreless region similar to the pars flaccida of land
mammals. The fibrous portion consists of a broad flattened triangular " ligament "
with one edge of its base attached to part of the tympanic ring, the base thus forming
part of the external surface of the " membrane." The attenuated apex of the
" ligament " formed by the convergent radial fibres is directed mesially and is
attached to a small process on the manubrium of the malleus. It is this ligament 1
which, in the opinion of the writers, is the true homologue of the fibrous portion
of the tympanic membrane of land mammals. The non-fibrous portion bridges
the space between the fibrous portion and the remainder of the tympanic ring.
In Phocaena phocoena (fide Bonninghaus) the non-fibrous portion merely forms a
few islets in the fibrous portion ; it is single and more extensive in area in the Globi-
cephala melaena specimen examined by the writers and in their opinion the " glove
finger " of the whalebone whales, described by Beauregard (1891) and Lillie (1910)
as well as by earlier anatomists, is the same structure very greatly enlarged.

The three auditory ossicles commonly found in the Mammalia are also represented
in the Cetacea. In the latter the malleus is fixed by the processus gracilis to the
margin of the tympanic annulus at the position of the Glaserian cleft. This con-
dition is not peculiar to cetaceans, Hyrtl (1845) and Bonninghaus (1904) having
described it in monkeys, carnivores and insectivores. The hammer-shape is main-
tained but it is the processus gracilis which resembles the handle. The head, in
addition to the portion recognized as such in other mammals, includes the manu-
brium, which is very much reduced in length to form a short tubercle. The processus
gracilis is longer and stouter than that of other mammals and in structure somewhat
resembles a short length of channel girder. At the junction of the malleus with
the Glaserian cleft the tympanic ring is developed into a strong wing-like lateral
projection buttressed to the processus gracilis. This projection known as the
sigmoid process is peculiar to the Cetacea, and plays an important part in the
mechanics of the cetaceans' hearing.

The head of the malleus is deeply notched by two large facets making a re-entrant
angle on its posterior aspect. Both of the facets have smooth, shallowly convex
surfaces covered with articular cartilage which with the corresponding facets on
the incus form part of a synovial joint. The radii of the convexities, as well as that
of the arcuate junction between the two facets, lie approximately at right angles
to the long axis of the tympanic ligament.

The tensor tympani muscle, which appears to be largely tendinous, is attached
to the ventral wall of the periotic near that part of the tympanic cavity associated
with the outlet of the Eustachian tube. Although in its course to the malleus it
is directed meso-laterally approximately in line with the long axis of the tympanic
ligament, the attachments of the muscle and tympanic ligament to the malleus are
displaced dorso-ventrally from one another by a small distance. The ligament and
muscle are not antagonistic to each other but form a mechanical couple.

The form of the incus is comparable with that of mammals generally. The facets
for articulation with the malleus are shallowly concave, furnished with articular

1 It will be referred to as the tympanic ligament in this paper.

HEARING IN CETACEANS 105

cartilages and meet at an angle which fits into the re-entrant angle of the malleolar
facets. The processus brevis is a short conical projection directed anteriorly.
The ligament attaching the incus to the periotic bone, the axis of the processus
brevis and the lateral margin of the processus gracilis make a common axis.

The stapes departs so little from the normal that its description is unnecessary,
but it should be mentioned that, contrary to the statements which have been made
that the foot is fused to the periotic at the fenestra ovalis, no evidence of this could
be found by the present writers. It is probable that the impression of fusion is due
to the perfect fit of the foot in the fenestra ovalis.

The tympanic bulla, the general form of which is sufficiently well known not to
require detailed description, may be said to have a mesial and a lateral portion
which meet ventrally in a roughly semi-circular ridge and are separated dorsally
by the cavity of the bulla. The mesial part is characterized by its stout petrous
nature (the bone in rorquals being up to 3-4 cm. thick), and by the smoothness
of most of its surface, both of these features are associated with its contiguity with
the air sinuses. The lateral portion is much thinner, and has a roughened outer
surface, these features being associated with its contiguity with the surrounding
fibrous layer, which is about 12 cm. thick in a large rorqual and strongly adherent
to the lateral surface of the bone. The bulla in the whalebone whales is attached
by two thin, flat pedicles placed respectively anteriorly and posteriorly and having
their planes approximately at right angles. The cavity of the bulla is continuous
with that of the middle ear and its associated air sinuses. In the Odontocetes the
anterior pedicle is absent and the posterior support does not involve bony fusion
of the tympano-periotic junction.

In all cetaceans the periotic is separated from the rest of the skull at the squamo-
mastoid boundary. In the rorquals the much attenuated mastoid process is loosely
wedged between the squamosal and basi-occipital bones — being maintained in
position by fibrous tissue. The mastoid process of the beaked whales and some
River Dolphins is less attenuated but much convoluted — the convolutions inter-
digitating with corresponding cavities on the postero-ventral tip of the squamosal.
There is no fusion of the bones.

In the Delphinidae the periotic is neither wedged into, nor interdigitated with,
the squamosal, but is separated from the bones of the cranium by an appreciable
gap. Between the mastoid process — which is very short — and the basi-occipital
there is a fibrous ligament which appears to be the only well defined point of attach-
ment to the skull.

The cochlea and semi-circular canals occupy only a relatively small part of the
petrous bone as a whole. It has been stated that in cetaceans the laminae spiralis
primaria and secondaria of the cochlea make a continuous bony septum across the
spiral labyrinth. The authors' dissections show that this is not so and that whilst
the gap between the two laminae is very narrow at the entrance to the scalae it
gradually widens towards their apex. There is a basilar membrane and organ of
Corti as in other mammals. In all but absolute size the cochlea conforms with the
usual pattern.

The cavity of the middle ear communicates with the Eustachian tube and with

106 HEARING IN CETACEANS

a system of pneumatic sinuses which in volume, extent and structure are peculiar
to the cetaceans.

In the region of the auditory organ accessory air sinuses occupy the space between
the mesial aspect of the periotic and the surrounding bones of the skull as well as
separating the periotic from soft parts on its lateral aspect.

Anterior to the tympanic bulla a very large air sinus is developed which in the
rorquals occupies nearly the whole of the pterygoid fossa. In the beaked whales
and dolphins this sinus is even more extensive and covers the greater part of the
base of the cranium in addition to the pterygoid fossa, which is very much enlarged.
In the Common Dolphin this sinus has an anterior extension which passes forwards
on the ventral surface of the rostrum for approximately two-thirds of the latter's
length.

The mesial walls of the pneumatic sinus are closely applied to the bones of the
skull. Laterally and externally the sinuses are closed by a tough, fibrous membrane
resembling periosteum to which some of the muscles of the lower jaw and palate
are attached. A comparative study of this membrane shows that in the more
primitive species it overlies thin sheets of bone and it has been suggested that the
enclosed sinuses are the homologues of the pneumatic sinuses found in the skulls
of terrestrial mammals (Monro, 1785).

Within the fibrous closing membrane there is a fibro- venous plexus which extends
over the whole of the inner surface of the air sinuses. The plexus is made up pre-
dominantly of large vessels which appear flattened when the sinuses are injected
with air or other media but become turgid when injected themselves under pressure.
In this state they obliterate the cavity of the air sinuses. Lining the latter and
covering the inner aspect of the fibro-venous plexus is a thick mucous membrane
which is continuous with that of the middle ear and Eustachian tube and is remark-
able for the richness of its glands and ducts. The openings of the ducts, which cover
the entire inner surface of the air sinuses, are less than • 1 mm. apart and lead into a
maze of smaller racemose channels and crypts lined with columnar epithelium and
goblet cells.

At the entrance to the ducts and on the exposed surface of the mucous membrane
is a layer of ciliated epithelium. The writers' observations, and those of Brazier
Howell (1930) on freshly killed specimens reveal that the sinuses are entirely filled
with an albuminous foam. Whether this is so in living animals cannot be ascer-
tained, but the anatomical evidence suggests that it is.*

The Eustachian tube diverges from the general air cavity at a point just anterior
to the tympanic bulla — and passes forwards for a short distance along the mesial
aspect of that part of the sinus which goes to occupy the pterygoid fossa : thereafter
it passes upwards and gradually narrowing, opens into the nasopharynx. Its
closing walls are made up of fibrous tissue and a fibro-venous plexus similar to, and
continuous with, that of the air sinuses. The lining mucous membrane is deeply
indented with valvular pockets and folds which are directed towards the choanae.

* Additional evidence of foam in the air sinuses has been received from Mr. D. E. Sergeant, Newfound-
land Fisheries Research Station, St. John's, Newfoundland, who writes (10-8-53): "Globicephala melaena,
female, 0/ 6" in length, dead about 1^- hours. Anterior region of sinus within pterygoid exposed by means
of saw cuts ; seen to be filled with foam".

HEARING IN CETACEANS

107

The fibro-venous plexus referred to above is supplied by an arterial plexus which
emerges from the external maxillary artery immediately anterior to the tympanic
bulla, and by small arterial branches which emerge from the same artery as it passes
forwards across the ventral surface of the cranium. The plexus is drained by three
distinct paths :

(1) By large vessels which communicate with the transverse and cavernous
sinuses of the cranium and drain eventually into the spinal-meningeal veins.

(2) By large vessels which join the external jugular vein via the vena
pterygoidea.

(3) By an intricate plexus of small veins which penetrates the fibrous
covering of the sinus at the angle formed by the lateral pterygoid and the
tensor palati muscles. This plexus is very dense and ramifies throughout
the mass of fatty tissue which lies on the mesial aspect of the lower jaw —
eventually coalescing into a single vessel which joins the external jugular
vein.

The aquatic environment of whales and the extreme narrowness of their external
auditory meatus have led most students of cetacean anatomy since Camper (1767)
to search for some mode of hearing other than that which is normal for terrestrial
mammals. Camper's hypothesis was to the effect that cetaceans could only hear
when lying quietly at the surface with the external aperture of the meatus out of
water. Hyrtl (1845), who recognized that whales must be able to hear well under
water, held the view that while the external meatus was unsuitable for the reception
of air-borne sound waves, it was suitable for the reception of " waves of compression "
such as are transmitted in water.

Buchanan's (1828) belief was that sound was conducted to the tympanic mem-
brane by way of the Eustachian tubes. This view was criticized by Claudius
(1858) on the ground that the Eustachian tubes are normally closed while the ceta-
cean is submerged. Claudius' own view was that the water-borne sound vibrations
were taken up by the bones of the skull, transmitted to the pneumatic sinuses,
thence to the auditory ossicles but principally to the membrane of the fenestra
rotundum which he regarded as the main oscillator of the cochlear fluid. Turner
(1913) agreed with Claudius' interpretation. Bonninghaus (1902) invoked the pinna-
like shape of the anterior portion of the tympanic bulla as a collector of sound waves ;
vibrations which he terms " molecular " are transmitted by way of the processus
gracilis through the ossicles to the fenestra ovalis.

Lillie (1910) says "the whale probably receives sound vibrations by means of
vibrating bony surfaces, after the manner of fishes. The tympanic bulla is a
relatively dense and heavy sounding-box, fastened to the periotic bone by two thin
pedicles, so that it could easily be set in vibration. The bulla is connected with the
fenestra ovalis by the chain of ossicles, the auditory apparatus being thus independent
of the tympanic membrane." Kellogg (1938) concurred with this view.

In the opinion of the present writers Hyrtl's interpretation is the most acceptable
one. The anatomical features of the cetacean ear, far from indicating that cetaceans
hear by abnormal means, support the view that the mode of hearing is essentially

108 HEARING IN CETACEANS

as in land mammals with precisely those modifications of structure which are
required for receiving water-borne vibrations.

Any working hypothesis relating to hearing in cetaceans must be based on the
assumption that these animals have the same requirements as terrestrial mammals.
These requirements are that hearing should be directional, discriminative, selective
and anticipative. The accepted hypotheses for directionality in hearing involve
an appreciation of an intensity and /or phase difference in the sounds received by
the two essential organs of hearing respectively. On the face of it it is difficult to
see how this can be achieved by bone conduction.

In the binaural reception of sound necessary for directionality the advantages of
the cochlea being situated at or near the extreme lateral limits of the skull are offset by
the likelihood of mechanical damage and extreme limitation of voluntary control-
ability of the sounds entering the external opening of the meatus. The muscularly
controlled pinnae of land mammals enables them, in addition to securing direction-
ality, to control the volume of sound received. In the Cetacea the anatomical
evidence indicates that binaural hearing is achieved by placing the two receptors
remote from the surface of the body, by isolating them acoustically from the rest
of the body and each other, and by connecting them to the surface of the body
laterally by an apparently acoustically efficient muscularly-controllable vibration
conduit. The acoustic isolation of the periotic is achieved by (i) its dissociation
from the adjacent bones of the skull, (2) the provision and maintenance of an air
gap between it and the rest of the skull at all hydrostatic pressures, (3) the foam
which, on the evidence of freshly-killed animals, appears to occupy the air sacs and
middle ear of the living animal. The first of these has already been described.
The second is achieved by the mechanism of the accessory air sinuses. As the
hydrostatic pressure increases with depth the blood is forced into the venous
plexus lying below the fibrous outer covering of the sacs. The volume of the air
sacs is correspondingly reduced and the compressed gas is contained in the rigid
tympanic cavity and in the space between the periotic and skull. With diminishing
depth the process is reversed. This interpretation of the function of the air sacs agrees
with that put forward by Beauregard (1891) except that the role of the venous plexus
does not coincide with his view.

Owing to the flattening of the large lining veins when the air sinsues are inflated
the blood drainage from the venous plexus and mucous membrane must be pre-
dominantly through the rich plexus of small vessels which passes through the mass
of adipose tissue occupying mesial concavities of the lower jaw. The very close
association of the plexus with the fatty tissue and the known high nitrogen absorp-
tion capacity of the latter compared with that of blood (6:1) suggest that the fat
may play a part in the absorption of nitrogen contained in the pneumatic sinuses.
That the mucous membrane of the sinuses may have a gas-secreting function is a
possibility which requires further investigation.

In considering the theory expressed by Claudius and implied by Bonninghaus
that sound waves are conveyed to the tympanic bulla through the air in the pneumatic
sinuses it must be pointed out that films of air surrounded by media of high acoustic
resistance like water do not transmit sound well. For example, a film of -i cm.

HEARING IN CETACEANS 109

of air surrounded by water reflects about 93% of the incident energy of a sound
wave of 1000 c.p.s. (Wood, 1941). If the air is surrounded by solid rock or presumably
by a substance equally hard, for example, petrous bone, the reflected energy is
about 99*93%. If in addition the air cavity is filled with foam the reduction in
energy would be even greater.

The efficiency of the bubbles in damping vibrations increases rapidly both as
their diameters and distance apart diminishes, thus in the cetacean as the animal
dives the acoustic isolation could be maintained in spite of the reduction in volume
of the foam-filled cavity.

In view of this apparent acoustic isolation of the essential organs of hearing it
was decided to examine the external meatus as a possible vibration conduit. Claudius
states that "All true cetaceans toothed as well as whalebone possess absolutely
no functioning ear passage " and subsequently after a very incomplete description
of the meatus states " The cord is then, neither homogeneous nor taut, and its
conducting power very slight, and by no means any greater than that of the sur-
rounding fat layer." The anatomical and experimental investigations carried out
by the writers do not support Claudius' conclusions, but point to a set of conditions
which are exactly the reverse.

The lining walls of the meatus appear extremely homogeneous (vide supra) and
the arrangement of the two most conspicuous auricular muscles would appear to be
precisely that necessary for maintaining either or both auditory conduits in any
state of tension required.

In order to obtain some indication of the sound conductivity of the meatus relative
to that of adjacent tissues a simple test was made. For the mechanical vibrations
required the output of a variable frequency oscillator (30-10,000 c/s) was connected
to the input of an amplifier of a loud speaker. A loop of dental floss was attached
to the centre of the loud speaker and the two ends attached to two small pieces
of perforated zinc which had been stitched to the wall of the meatus and the
adjacent fibrous tissue respectively. The specimen used consisted of a large piece
of the squamo-occipital region of the head of a Fin Whale which, deep frozen
while fresh, had been defrosted and dissected to expose the tympanic cavity.

For indicating the relative strength of received vibration signals a gramophone
pick-up was connected to an amplifier, the output of which was connected to a
rectifying millivoltmeter.

With a fixed output from the oscillator and with the pick-up needle placed at
the external end of the meatus, the input potentiometer of the amplifier in the pick-up
circuit was adjusted to give a reading of 0-4 mv. at the millivoltmeter. To obtain
the same reading with the pick-up needle placed at the malleolar end of the tympanic
ligament the input potentiometer of the amplifier had to be adjusted. This adjust-
ment corresponded to a voltage ratio of 10 : 1 so that assuming that the voltage
reading is directly proportional to the mechanical vibration picked up, the signal
received at the malleolar end was about 1/10 that applied at the outer end. When
attempts were made to pick up signals on the periotic bone, on adjacent skull bones,
on the muscular or on the fibrous tissue in the neighbourhood of the tympanic cavity
the millivoltmeter reading fell to zero. With the apparatus available for this simple

no HEARING IN CETACEANS

test the limits set by the oscillator output and the amplification in the pick-up
circuit, prevented a quantitative comparison being made between these last tests
and the first one. The indication is, however, that the signal in any case is less than
a quarter of that obtained at the malleolar end of the tympanic ligament.

The results confirmed earlier preliminary tests made with a stethoscope. Because
of the mechanical difficulty of placing the stethoscope on the malleus the vibration
was applied to the point of attachment of the tympanic ligament and the signal
picked up at the external end of the meatus. With this arrangement good signals
were received, but with the stethoscope applied to surrounding tissues results were
negative.

To satisfy the discriminative condition it must be assumed that the cetacean
cochlea is as selective as that of terrestrial mammals. Certainly its general anatomy
suggests that this is so. The physical properties of water-borne vibrations, however,
differ markedly from those which are airborne. The pressure-amplitude for the
same intensity and frequency of water and air-borne sound is in the ratio 61 : i,
and the displacement amplitude i : 61. The mathematical treatment of the subject
is not within the scope of this paper but it can be shown that adjustments of ampli-
tude and pressure to values normally experienced in the cochlea by terrestrial mammals
are probably achieved in cetaceans by the modifications of the malleus and incus.

Before describing the functioning of the cetacean ossicles it is necessary to repeat
the often quoted modus operandi of the human ear. The following is from Gray's
Anatomy (1946) :

" The handle of the malleus follows all the movements of the tympanic membrane, while
the malleus and incus rotate together about an axis which runs through the short process of
the incus and the anterior ligament (and therefore the processus gracilis) of the malleus. When
the tympanic membrane and handle of the malleus move inwards the long process of the incus
also moves in the same direction and pushes the base of the stapes towards the labyrinth."

A mechanical study of their ossicles shows that the foregoing description is true for
cetaceans with one important difference. In man and other terrestrial mammals
the long axis of the malleus is parallel with the plane of the tympanic membrane
and the manubrium is firmly attached to the latter as far as its centre. Thus the
amount of movement of the centre of the tympanic membrane is identical with that
of the tip of the manubrium. In cetaceans the membrane (i.e., the tympanic
ligament) is attached only to a point at the tip of the manubrium of the malleus.
The attachment to the tip is so situated that a line through the centre of the tympanic
ligament is parallel with the long axis of the manubrium. This means that any
small movement of the meatal end of the tympanic ligament produces a relatively
large movement of the tip of the manubrium and therefore of the tip of the long
process of the incus. In order that this amplification be produced the effective
movement of the malleus must be solely rotational and this is achieved by the stout
channel-girder construction of the processus gracilis and its firm fusion to the tym-
panic ring. In addition, its lateral edge is strongly buttressed by the sigmoid
process of the tympanic ring. This construction ensures that lateral movement
of the malleus is obviated and that the processus gracilis is capable of torsional
vibration only. The angular articulation of the malleus with the incus produces

HEARING IN CETACEANS in

a condition in which there is positive oscillatory drive by rotation only. The smooth
curved surfaces of the facets tend to slip over one another in one or both directions
when the oscillation is transverse or longitudinal to the processus gracilis. This
factor must be of great importance in the exclusion of extraneous vibrations. In
this last connection it is necessary to examine the hypothesis of Lillie and Kellogg
(v. supra) in more detail and as a preliminary it will be useful to recall Beauregard's
(1894) conclusions on the subject.

" The tympanic bulla cannot be regarded as an apparatus for resonance as has been asserted
by various anatomists and physiologists (Hunter, Home, Esser, etc.). The massiveness of the
walls of the bulla and its shape recalling more or less that of a shell have evidently given birth
to this idea, but the facts do not justify it. These so-called resonators are more or less completely-
enveloped by soft parts which do not favour their action. The shape is extraordinarily variable
in reality and the massiveness of the walls is necessary to obtain a rapid equilibrium ol pressure
between the air which fills the bulla and that of the exterior. Savart was therefore correct in
saying that the tympanic cavity has the function of retaining near the entrance of the labyrinth
and on the internal aspect of the tympanic membrane an atmosphere of which the physical
properties are constant."

In agreeing with this view the writers would draw attention to additional
anatomical details which support it. (1) the great thickness and strength of attach-
ment of the fibrous capsule which envelops the ventro-lateral aspect of the bulla,
this being the part which would be most easily thrown into vibration ; (2) the extreme
density and thickness of that part of the bulla which is exposed to air on both
surfaces, the internal damping of such a mass must be considerable ; (3) the angular
set of the two thin pedicles is such that the bulla would not easily be thrown into
vibration, but assuming it were it would be expected to vibrate about an axis running
through the pedicles. Examination of the positioning of the articular facets of the
malleus in relation to this axis indicates that no drive to the incus could be achieved
(the connection between the malleus and incus has already been described).

The majority of sounds in water can be regarded as being conveyed from a point
source by spherical pressure waves. When such waves impinge on the cetacean
body, and assuming the least favourable acoustic conditions, i.e., that the general
shape and various tissues of the body are non-reflecting, the body mass would form
no barrier to the progress of the waves except in the neighbourhood of the ear.
Here the sound energy, except that conveyed by the external meatus, must undergo
almost complete reflection because of the gaseous content of the tympanic cavity
and accessory air sinuses. The walls of the external meatus would be thrown into
longitudinal vibration whether or not a continuous free lumen existed. Such
vibrations would be conveyed to a point on the mesial end of the manubrium mallei
by the convergent fibres of the tympanic membrane. The oscillations of the malleus
thus initiated would be conveyed to the cochlear fluid over the bridge formed by the
auditory ossicles.

In terrestrial mammals the intensity of sounds received can be governed by
orientation of the head, but principally by the use of auricular muscles. In the
Cetacea the orientation of the head in relation to the rest of the body is limited by
the absence of a well-defined neck. If, however, it is accepted that the sounds

ii2 HEARING IN CETACEANS

received at the cochlea of the cetacean are conducted along the external auditory
meatus by longitudinal waves the degree of tension of the walls of the meatus must
be a controlling factor in sound intensity. The m. auricularis occipitalis profundus
and the m. zygomatico-auricularis of the Porpoise described by Bonninghaus seem
particularly well suited in position and structure for the function of adjusting the
tension. The meatus, however, is attached to the tympanic annulus and an addi-
tional mechanism is required to maintain the tension of the fibrous portion of the
tympanic membrane lying within the tympanic cavity. The positioning of the
non-fibrous portion of the membrane in relation to the fibrous portion is such that
under the influence of the gaseous pressure within the tympanic cavity the necessary
tension could be provided. The final expression of this function can be seen in the
" glove-finger " of the rorquals. In them the non-fibrous portion is enlarged in
correspondence with the greatly elongated tympanic ligament.

The arrangements which exist in terrestrial mammals for making small temporary
adjustments to the tension of the tympanic membrane are also present in the Cetacea,
but in view of the coarse adjustment required it is doubtful whether the tensor
tympani muscle is as effective in this respect as it is in terrestrial mammals. The
palatal and lateral pterygoid muscles are however very well developed in cetaceans
and their relationships to the Eustachian tube and pneumatic sinuses suggest that
they may play an important part in the adjustment of the gaseous pressure within
the tympanic cavity and consequently of the tension on the tympanic ligament.

BIBLIOGRAPHY
Beauregard, H. 1894. Recherches sur l'appareil auditif chez les Mammiferes. J. Anat.,

Paris, 1894 : 367-413, 3 pis.
Bonninghaus, G. 1904. Das Ohr des Zahnwales, zugleich ein Beitrag zur Theorie dcr

Schalleitung. Zool. Jb., 19 : 189-360, 28 figs., 3 pis.
Buchanan, T. 1828. Physiological Illustrations of the Organ of Hearing. London.
Camper, P. 1767. Verhandeling over het Gehoor van den Cachelot, of Pot-Walvisch. Verh.

Holland. Maatsch., 9 (3) : 193-229, figs.
Carte, E., and Macalister, A. 1868. On the Anatomy of Balaenoptera rostrata. Phil.

Trans., 158 : 201-261, 5 pis.
Claudius, M. 1858. Physiologische Bemerkungen iiber das Gehororgan dcr Cetaceen und das

Labyrinth der Sdugethiere. Kiel.
Home, E. 1812. An Account of some Peculiarities in the Structure of the Organ of Hearing

in Balaena mysticetus of Linnaeus. Phil. Trans., 102 : 83-88, 2 pis.
Howell, A. B. 1930. Aquatic Mammals : their Adaptations to Life in the Water. Pp. vii ;

338. Springfield, Illinois.
Hunter, J. 1787. Observations on the Structure and Oeconomy of Whales. Phil. Trans.,

77 : 371-450, 8 pis.
Hyrtl, J. 1845. Vergleichend-anatomische U ' ntersuchungen iiber das inner e Gehororgan des

Menschen und der Sdugethiere. Pp. viii, 139, 9 pis. Prag.
Johnston, T. B., and Whillis, J. (Editors). 1946. Gray's Anatomy, 29th ed. 1-1597, figs.

London .
Kellogg, R. 1938. Adaptation of Structure to Function in Whales. Publ. Carneg. Instn.,

501 : 649-682.
Lillie, D. G. 1 910. Observations on the Anatomy and General Biology of some members

of the Larger Cetacea. Proc. Zool. Soc. Lond., 1910 : 769-792, 1 pi., 10 figs.
Monro, A. 1785. The Structure and Physiology of Fishes. 1-128, 44 pis. Edinburgh.

HEARING IN CETACEANS 113

Turner, W. 1914. Observations on the Auditory Organ in the Cetacea. Proc. Roy. Soc.

Edirib., 34 : 10-26, 4 figs.
Wood, A. B. 194 1. A Text-book of Sound, being an account of the Physics of Vibrations with

special reference to recent theoretical and technical Developments. 1-578, figs., tables and plates.

London.

Munesato Yamada's " Contribution to the Anatomy of the Organ of Hearing of Whales,"
reprinted from the Scientific Reports of the Whales Research Institute, No. 8, 1953, was received
after completion of the foregoing preliminary account. It will be considered in the more
detailed paper which the authors have in preparation.

EXPLANATION OF PLATES

PLATE 4
Fig. i. — Dissection to expose the ventral aspect of the left ear of a Pilot Whale, Globicephala
melaena. Cut surfaces of bone hatched and the corpus cavernosum shown reflected. Abbre-
viations used in this and the following figures : —

— Accessory air sinus.

— Auricular cartilage.

— Anterior ligament.

• — Anterior pedicle.

—Blubber.

■ — Cochlea.

— Corpus cavernosum.

External auditory meatus.

AAS

EAM

fo — Fenestra ovalis.

fr — Fenestra rotundum.

fvp — Fibro-venous plexus.

gl — Glove finger.

i — Incus.

li — Ligamentum incudis.

m — Malleus.

-Manubrium mallei.
-Occipital condyle.
-Periotic.
-Processus gracilis.

MM
OC
P
PG

pp — Posterior pedicle.

pts — Pterygoid sinus.

s — Stapes.

sc — Semicircular canal.

sm — Stapedial muscle.

sp — Sigmoid process.

ta — Tympanic annulus.

tb — Tympanic bulla.

tl — Tympanic ligament.

tt — Tensor tympani muscle.

Fig. 2. — Detail of figure I.

PLATE 5

Fig. i. — Dissection to expose the ventral aspect of the left ear of a foetal Lesser Rorqual,

Balaenoptera acutorostrata. (Abbreviations as for Plate 4, figure 1.) Corpus cavernosum shown
reflected.

Fig. 2. — Right tympanic annulus and auditory ossicles of a Humpback Whale, Megaptera
novaeangliae. The dotted line indicates the axis of the tympanic ligament. (Abbreviations
as for Plate 4, figure 1.)

JAN 1954

Bull. B.M. {N.H.) Zool. 2, 5

Fig. 1.

EAM

Fig. 2.

Bull. B.M. {N.H.) Zool. 2, 5.

FVP

Fig. 1.

Fig. 2.

PRINTED IN GREAT BRITAIN BY
AD LARD AND SON, LIMITED,
BARTHOLOMEW PRESS, DORKING

PRESENTED

5 OCT 1954

THE "ROSAURA
EXPEDITION

BULLETIN OF
THE BRITISH MUSEUM (NATURAL HISTORY)
ZOOLOGY Vol. 2 No. 6

LONDON: 1954

THE "ROSAURA" EXPEDITION

Pp. 115-239; Pis. 6-9; 40 Text-figures

BULLETIN OF

THE BRITISH MUSEUM (NATURAL HISTORY)

ZOOLOGY Vol. 2 No. 6

LONDON : 1954

THE BULLETIN OF THE BRITISH MUSEUM
(NATURAL HISTORY) instituted in 1949, is
issued in five series corresponding to the Departments
of the Museum, and an Historical Series.

Parts appear at irregular intervals as they become
ready. Volumes will contain about three or four
hundred pages, and will not necessarily be completed
within one calendar year.

This paper is Vol. 2, No. 6 of the Zoological
series.

PRINTED BY ORDER OF THE TRUSTEES OF
THE BRITISH MUSEUM

Issued September, 1954

Price Thirty Shillings

THE "ROSAURA" EXPEDITION
1937-1938

CONTENTS

Page

1. Gear, Narrative and Station List. J. S. Colman . . . .119

2. Under- Water Illumination and Ecology in Tropical Estuaries. J. S.

Colman and L. H. N. Cooper ....... 131

3. The Echinodermata. D. D. John and A. M. Clark .... 139

4. Fishes. Part I : Families Carcharhiniidae, Torpedinidae, Rosauridae

(nov.), Salmonidae, Alepocephalidae, Searsidae, Clupeidae. Denys

W. Tucker 163

5. Sponges. M. Burton . . . . . . . . .215

FOREWORD

The collections made by Mr. J. S. Colman during the cruise of Lord Moyne's yacht " Rosaura "
in the north and central Atlantic and the Caribbean region were received at the British Museum
(Natural History) only a short time before the outbreak of war. In consequence little work
was done on them at the time, and during the post-war years there have been many tasks of
greater urgency. In some groups the material collected adds little to existing knowledge
and no special report on it, as a whole, is warranted. In such instances new knowledge has
been, or will be published elsewhere ; published papers dealing with " Rosaura " material are
listed by Mr. Colman in the bibliography attached to his narrative of the cruise (p. 130).

The present Bulletin contains a random collection of reports at present available. Others
may be expected from time to time in this Bulletin.

i. GEAR, NARRATIVE AND STATION LIST

By JOHN S. COLMAN.

In 1937 Lord Moyne took his yacht " Rosaura " on a long cruise in the North Atlan-
tic and decided to add deep-sea collecting to the other activities of the trip. I was
fortunate to be taken on as zoologist.

We left Southampton on 21st August, 1937, and returned there on 6th February,
1938, after covering 18,565 miles. We called at the Scilly Isles on the way out,
and then proceeded straight to Angmagssalik in East Greenland, where we spent
four days. We then rounded Cape Farewell to Julianehaab, the capital of southwest
Greenland, and visited several fjords in the neighbourhood. Next we had to double
back to St. John's, Newfoundland, in order to fit new propellers in the dry-dock
there, as ice had removed a blade from the port propeller in Angmagssalik harbour.
This prevented a projected run up the west coast of Greenland to Disko, so instead
we went from St. John's to the Labrador coast. In southern Labrador the harbours
are not good, as the low, smooth hills afford very little protection from the wind,
and the bottom is often hard and poor holding-ground ; we dragged our anchors in
two successive " harbours " and had to put to sea each time. After Labrador we
visited three of the fine inlets on the west coast of Newfoundland and then proceeded
to Cape Breton Island, where " Rosaura " was again forced to go out to sea when the
wind freshened ; this, however, was the last heavy weather we had on the entire
cruise.

The first part of the voyage ended at New York ; our route was then
via Charleston, Nassau, Cap Haitien in Haiti, Jamaica, Grand Cayman, Swan
Islands and the Bay Islands of Honduras, to Belize in British Honduras. The Bay
Islands and Swan Islands turned out to be unexpectedly interesting, the latter bio-
logically and the former by providing enormous quantities of pottery which is pro-
bably pre-Columbian, so we revisited them before running south to Colon on the
Panama Canal. From here the course was by the San Bias Archipelago of Panama
to Puerto Cabello and La Guayra in Venezuela, then by Grenada, Tobago, Little
Tobago and Trinidad to Demerara and the Barima River, and on to Para. Christmas
was spent in Pernambuco, after which a visit was paid to St. Paul's Rocks on the
way across the Atlantic to the Gambia River, up which we went for 120 miles. We
then had an exasperating three weeks in Dakar with two visits to the dry-dock,
but we thereby missed two severe January gales off the English Channel, and our-
selves had almost a glass calm from Finistere to Southampton. The whole cruise
was described by Lord Moyne in Atlantic Circle (Blackie 1938).

" Rosaura " was a vessel of some 1,500 tons, 282 feet long and drawing about
12 feet of water. (She was lost during the 1939-1945 war). She started life as
the cross-channel steamer " Dieppe " and was converted from steam turbines to

ZOOL. 2, 6 7

120 THE "ROSAURA" EXPEDITION

diesel engines, but the lighting, heating and capstans were still worked by steam.
I mention this because the chief piece of collecting equipment was the steam winch
off R.R.S. " Discovery I," which could thus be built into " Rosaura " with the mini-
mum of alteration. This winch had some 3,500 fathoms of wire and did its work
splendidly, giving very little trouble. The only place for it was on the deck just
forward of the bridge, and this proved to be the best possible position, for two
reasons. First of all, the most practicable lead for the wire was then over the bows,
and this meant towing with the engines going astern ; going slow ahead on one
engine " Rosaura' s " minimum speed was 3 or 4 knots, but going astern this could be
reduced to 2 knots, about right for plankton work or for dredging. Secondly,
" Rosaura " had a very strong idiosyncracy, that of always trying to lie stern to
wind ; this made towing over the bows a simple matter, and sometimes neither the
wheel nor the engine-room telegraph had to be touched during the hour or more that
the net was towing.

For work on the bottom we had a 10-foot Agassiz trawl and dredges 2 and 3 feet
across. The Agassiz trawl was on the whole very successful and only came to
grief once, when we tried a trawl in 1,000 fathoms off Swan Islands ; about 70
fathoms of wire came up in a knot with the cod-end of the net in the middle ; the
deepest successful haul was from 700 fathoms near the Canaries.

Occasions for deep bottom trawling were limited by the scarcity of suitable depths
(1,000 fathoms was about our maximum), so the bulk of the collection consists of
deep-sea plankton from between about 700 fathoms and the surface, caught in a
stramin net with a diameter of 2 metres. Since we did not have closing apparatus,
the net was sent down and hauled in open. It was towed horizontally for an hour,
or sometimes more, in deep water, and then took about half an hour to heave in,
so the catches must contain a good deal of material from the upper layers and the
surface. In addition I made several coastwise and estuarine hauls with a small
silk tow-net 50 cm. in diameter.

All the nets and frames were obtained from the Plymouth laboratory of the Marine
Biological Association, where they were made exceedingly well and with remarkable
speed, and the British Museum supplied jars, tubes, formalin and alcohol, of which
last I used 45 gallons. Every haul was sorted and bottled off at once unless the
ship was rolling too heavily. A plankton haul would take a day and a half and a
dredged haul from one to three days, as the sorting was done in fair detail, particu-
larly in the more abundant groups such as mollusca and Crustacea.

It was too rough to do any work on the way over to Greenland, so the first collecting
was done in Angmagssalik harbour. The bottom is rather like the surrounding
mountains, which are steep and rugged, and the harbour has not been charted, so
we dredged from a motor-boat close to the rocks where we could find bottom in
20-30 fathoms. Although the depth was altering constantly the bottom was of
mud, at any rate in places, and it was extraordinarily rich in life. There were 14
species of bivalves, about as many of Polychaetes, many algae and many Crustacea,
but curiously enough no crabs. Crabs seem to be rare in Greenland, for in five
dredge-hauls and shore-collections I caught only two in all. Otherwise the Green-
land marine fauna and flora are both remarkably rich, but are on the whole not

THE "ROSAURA" EXPEDITION 121

easy to come at. In places like Angmagssalik harbour the sea floor is extremely
uneven, so much so that one night when " Rosaura " was shifting her moorings
during a gale the echo-sounding gear showed every depth from 14 to 70 fathoms
and back again in under a quarter of a mile ; then, again, over the coastal shelf the
sea-room necessary for fairly deep trawling is strictly limited by the proximity of
icebergs, and the bottom is so foul with glacial boulders that a long haul is bound
to result in the loss of some part of the gear. A canvas sock outside the net is
absolutely essential. Our one successful bottom-haul off Greenland, Station 6,
brought up large numbers of brachiopods, bivalves, polychaetes and ophiuroids ;
the coastal plankton also contained many ophioplutei, so that brittle-stars may
be a dominant group in these parts.

On leaving Greenland we made for St. John's, Newfoundland, to fit spare pro-
pellers as the ice had broken a blade off the port propeller while we were anchored
at Angmagssalik. On the way we tried out the 2-metre stramin net for the first
time, and rather rashly did so at night. It was a most instructive station, as nearly
everything happened except the loss of the net or of the catch. We could not use
the port engine, the sea was not running true with the wind and " Rosaura " accord-
ingly lay rather across the sea, rolling heavily, and to crown all the West Greenland
current kept driving the ship over the wire. Sometimes the wire paid off ahead as
it should, sometimes it grew straight up and down, and twice it went right under the
keel and paid off astern on the starboard side. Finally a davit-guy parted but this
was countered with a jury tackle from the anchor davit, and eventually the net was
brought aboard. Rather to our surprise it contained a rich and very phosphorescent
catch, including some eight or ten species of scypho-medusae — more than at any
other station. This catch was more phosphorescent than any other, probably
because it was the only deep plankton haul made where the surface water was cold.

In St. John's harbour there were many patches of red water which gave quite a
distinctive appearance to the inlet. They were caused by myriads of cladocera,
each containing a yellow drop of oil.

St. John's welcomed us with an unprecedented heatwave at 92 F., but after we
left for Labrador we were told the same thing wherever we went for a fortnight —
H What a pity you weren't here two days ago ; this is the first foul weather we have
had for months !" This weather limited collecting opportunities on the Labrador
coast, and we could not even remain in harbour, as I have mentioned, since the hold-
ing-ground was so poor. I did get in a dredge haul from the dinghy between two
gales (Station 10), and so found out what the charts mean by describing the bottom
as " coral " over so much of this coast. The net came up entirely filled with a
branching coralline alga, which evidently covers large areas of a rocky bottom
and is itself very hard. It acts like roller-bearings under the flukes of a boat's anchor
and must make the holding-ground even worse than bare rock.

We came south down the west coast of Newfoundland and tried out the Agassiz
trawl in Bonne Bay. This is an inlet with only 6 fathoms at the mouth, but with
120 fathoms further in where we trawled ; there is no similar depth nearer than
forty miles out in the Gulf of St. Lawrence. The bottom of Bonne Bay is of mud,
with some stones and a good deal of waterlogged lumber. More than half the animals

122 THE "ROSAURA" EXPEDITION

were ophiuroids in prodigious numbers, but most of the main groups were represented
and the collection was unusually varied for such a muddy bottom. Three large scar-
let anemones were rather unexpected, but these appear to have a foot specially
adapted for living in mud ; the lower end of the body is solid and hemispherical,
and surrounded by a wide and muscular frill ; the anemones came up with the space
between the frill and the rounded foot filled with compressed mud, and it seems that
the anemone can use this as a more or less solid platform on which to stand. They
were problably about 6 inches long when full expanded.

In Liverpool Bay, Nova Scotia, there was a curiously barren beach. The off-
lying rocks were densely covered with algae and carried thousands of whelks and
sea-urchins, but there was a swell running which made it impossible to land on them.
The beach referred to looked like firm brown sand, but was spongy to walk on and
turned out to be a peat made of sawdust and chips from a small pulp-mill more than
a mile up the bay. This beach was quite extensive, over a foot thick, and seemed
to be permanent and it was quite a distinct geographical feature. As the pulp-
mill was still small and had been in existence only a few years, it will be interesting
to see how this beach develops. It is remarkably barren, containing neither poly-
chaetes, amphipods, nor any other metazoa that I could find.

After leaving New York we went to the Bahamas and spent the next two months
in the West Indies, Central America, and the north coast of South America. The
weather was ideal for marine collecting, with gentle trade-winds and calms, and
good visibility for navigation ; no hurricanes came anywhere near us. We made
a number of mid-water plankton hauls in deep water, dredged when the bottom was
within reach, and at Grand Cayman, Swan Islands, Bay Islands and British Honduras
I carried out shallow-water dredging and shore collecting.

The first really interesting catch was a live Spirula, which was brought up at
Station 15 between Jamaica and Cuba. When the " Challenger " caught its single
specimen of Spirula in a dredge they naturally thought it was a bottom living animal,
but all subsequent evidence goes to show that it lives a planktonic or nektonic life
some way below the surface. The " Challenger " specimen may have been caught
by the dredge on its way from the bottom to the surface while being hove in. Spirula
shells occur sometimes in such vast numbers that they form white beaches extending
for miles, but until fairly recently the animals themsleves have been very rare in
collections. The Cambridge Natural History of 1903 says that there were then only
five specimens known, but since then the " Michael Sars " has caught several and
the " Dana " about one hundred, so its capture is no longer the event it used to be.
Later on we caught three more, all very young. It was interesting to watch the
animal alive for some time. Unlike many other cephalopods, such as Sepia, it
does not lie horizontally in the water but always head down. It swims up and down
with its fins, rather slowly, but can also propel itself violently tail first by squirting
water from its syphon. (I never saw one eject any ink.) Sometimes it would
hang on to the side of the jar for a few seconds, and if one put one's finger in the
jar the Spirula would at once cling on and nibble gently. In colour the body was
white, and curiously rough ; this part looked exactly like the ash of a cigar. The
tentacles were reddish-brown, the head blue-gray, and the mantle-edge and hinder

THE "ROSAURA" EXPEDITION 123

end reddish brown. Right at the after end between the fins there is a round disc,
which the older books describe as a sucker. Schmidt (1922), however, kept Spirula
alive for some days in the " Dana," and found that this disk is a luminescent organ
which gives off a steady but dim light (see also Bruun, 1943). I did not
see this myself.

Spirula 's habit of swimming constantly with its hind end upper-most may explain
why it it is not more often caught. If it is startled by the bridle of the net coming
through the water and gives a convulsive leap, it will automatically move several
feet straight upwards and the net will pass beneath it. An animal like a prawn such
as Acanthephyra, which swins horizontally, will shoot off in a horizontal direction
when similarly startled and may sometimes go straight into the approaching net.

Just north of Turneffe Island, British Honduras, we used the Agassiz trawl in
500 fathoms. The bottom ranged from fine mud to gravel and was bright yellow
in colour. The fauna was rich and included large numbers of Stephanotrochus dia-
dema, a deep-sea coral, which is a solitary form like Fungia, and is unattached when
adult. The tissue is a very deep purplish brown, almost black, and it is disappoint-
ing that none of them would expand even in cold water in the dark after being
brought on board. Stephanotrochus is in some ways a rather unwelcome catch ;
near the edge there are twelve very sharp vertical plates standing up a quarter of
an inch from the body of the coral, which is itself an inch and a half across ; these
plates cut up into ribbons such animals as prawns and polychaetes.

While we were in this region we visited the Bay Islands of Honduras three times.
They are high islands, separated from the mainland thirty miles away by water
several hundred fathoms deep. The population are mostly English-speaking, as
the islands were an informal British colony until a hundred years ago. The most
interesting find was enormous quantities of pottery of uncertain age and affiliations
which cannot be dealt with here, but in addition to this a gecko turned up in a sackful
of lizards which had been bought as food for snakes. According to Parker (1940)
this gecko is a new species of Sphaerodactylus, a genus which extends all over the
Central American and West Indian regions and contains many species, some of
which are peculiar to single islands. The Antillean species are distinguished by
possessing a groove down the back with very small scales, a feature which is mis-
sing in the mainland forms. This Bay Island gecko {Sphaerodactylus rosaurae
Parker), though so close to the mainland of Honduras, has the groove and small
scales, and in this way resembles species living 2,000 miles further east.

We paid three visits to the Bay Islands and two to Swan Islands. The latter
are two small coral islands, remarkably interesting to a biologist. They lie about
150 miles off the north-east corner of Honduras, and are surrounded by water of
1,000 fathoms and over ; each island is about a mile and a half long by a mile wide,
is composed chiefly of coral limestone, and is densely forested. The western island
has a landing-beach and is inhabited by 21 Cayman islanders who grow coco-nuts.
They have cattle on the island and rather surprisingly were able to sell us a sheep.
This island was used as a careening place by one of the later pirates, Captain Swan,
who gave his name to it. There is a large kitchen-midden on the western island
containing crude pottery and burnt shells, which shows that Indians once lived there.

I2 4 THE "ROSAURA" EXPEDITION

The eastern island has no beach and is surrounded by cliffs of 30 to 40 feet, very
rough and sharp, which make landing difficult ; it has apparently never been inhabi-
ted, even by Indians. It is thickly wooded and has a remarkable fauna, including
earthworms, millipedes, centipedes, at least eight species of land gastropods, an
iguana, a lizard, a snake, many species of bird, and a mammal peculiar to the island,
Capromys thoracatus. This is a kind of cavy, about the size of a muskrat but with
very coarse fur. It must have as small a range as any mammal ; the other species
of the genus are confined to Cuba and Jamaica.

In Stations 28 (off the Bay Islands) and 42 (off N.E. Brazil) we caught in the
stramin net 3 and 2 specimens respectively, all young, of the extraordinary deep-sea
cephalopod V ampyroteuthis infernalis, of which less than a hundred specimens
appear to exist. These larvae have been described in detail by Pickford (1939).

Another catch of note was off Grenada in Station 34. Most of the Antilles are
very steep-to and surrounded by water too deep for us to dredge in, but three miles
off St. George, Grenada, there is a shelf at 450 fathoms on which we towed the Agassiz
trawl. It fouled the bottom after 15 minutes, bringing the ship up and bending
the trawl frame, but the wire held and the net was undamaged. The bottom appears
to be covered with a finely-branching primnoid alcyonarian, Stenella, and a singu-
larly beautiful sponge, Aphrocallistes. About two bushels of these were in the net,
and among them was a remarkable lobster, Phoberus caecus, which is one of the
largest known deep-sea Crustacea. It was pink, with vestigial eyes and very long
chelipeds armed with long and very sharp teeth ; the carapace was spiny and rather
thin, while the antennae were like wire and over 2 feet in length. As far as I know,
only two individuals of this species have been caught before, both by the " Blake "
in 1878 on the same grounds ; we were within a mile of the " Blake's " position.

Off the River Orinoco we used the ship's otter trawl three times in order to get
some food-fish for the larder. In this we had no luck, but the trawl brought up a
number of interesting invertebrates, including the only crinoids caught on the trip
and a very fine basket star-fish, Astrophyton muricatum. The latter was contracted
into a tight ball as big as one's fist but expanded nicely in a bath. It was narcotized
successfully with menthol and pickled at its fullest extent. The length of one of
the long arms was 47 cm., and the whole animal when extended was more than
a yard across.

The rivers of northern British Guiana form an anastomosing series of tidal streams
with remarkably turbid water. I collected samples of the water in the Barima and
Demerara Rivers, and these have been examined by Dr. L. H. N. Cooper for their
light penetration values. These values are so low that it is difficult to understand
how any effective photosynthesis can take place in these streams ; nevertheless
the animal population is both abundant and varied, including microphagous feeders
such as copepods and even bottom-living barnacles, and in the Demerara there were
enormous numbers of drifting algae. These water-samples, with some others collec-
ted for comparison in the Gambia River, will be discussed by Dr. Cooper and myself
in a subsequent paper in this series.

Off northern Brazil we got remarkable catches of pelagic molluscs, including
pteropods, heteropods, squids and octopus, and at one station (41) the stramin net

THE "ROSAURA" EXPEDITION 125

distinguished itself by catching a 7-foot blue shark, Carcharinus sp. The shark,
an immature female, came in dead, but was accompanied by four live remoras. The
shark's stomach was quite empty.

To a marine biologist perhaps the most interesting part of the cruise was the
crossing of the Central Atlantic from Pernambuco to the Gambia River. Three
deep plankton hauls were taken (Stations 43, 45 and 46), in different bodies of water.
The two southern stations, in the Equatorial current, were distinguished from all
the other midwater hauls in producing neither red Sagittae nor red prawns such as
Acanthephyra. Station 46 was in the east-going Guinea current, where Acanthephyra
and red Sagittae reappeared.

Several remarkable animals were found in these stations. Three squids in parti-
cular were unlike any others we caught. One, Calliteuthis hoylei, was rufous brown,
with markedly retractile and protrusible eyes, and was scattered all over with power-
ful light organs equipped with lenses ; it must be a brilliant object when illuminated.
The other two both had crumpled bag-like bodies, translucent and colourless, and
belonged to the group Cranchiidae. Each had two long tentacles dotted with
chromatophores, but whereas one had arms of moderate length and very prominent
but unstalked eyes, the other had excessively small arms and huge eyes on stalks ;
this last has been identified as Bathothauma lyromma. Chun (1910, 1915) figures
similar forms tightly blown up and consequently streamlined, but our specimens were
half deflated while still alive. It is hard to believe that they are very active creatures.

Amphipods were specially numerous in the Central Atlantic, and included several
Phronimae which actively paddled their cases about for some time after being
caught ; but the finest was a perfect specimen of Cystisoma about 5 inches in length,
colourless except for a yellow stomach and a very faint pink tinge to the legs. It
was so transparent that in water the stomach was the only part to be seen, not even
the brain and nerves becoming visible until after the animal had been killed and left
in formalin for some time. The eyes were enormous, covering the whole dorsal
surface of the head, but were colourless and as transparent as the rest of the body.
It is hard to understand how these animals can see at all clearly.

An interesting visit was paid to St. Paul's Rocks, which are right in the middle of
the Atlantic Ocean, less than a degree north of the Equator. The rocks are only
60 feet high at the most and the most southerly point is less that 300 yards from the
most northerly ; probably no part is as much as 10 yards from the sea. This tiny
landmass is apparently not of volcanic origin, but is formed of plu tonic rocks ; it
can with some justice be considered as the world's smallest continent. The rocks
stand in 20 or 30 fathoms on a shelf only about half a mile wide, which is surrounded
on all sides by water more than 2,000 fathoms deep. As far as we could make out
from our echo-sounding gear the sides of the shelf are virtually perpendicular, a
feature reminiscent of coral-reefs. No coral, however, is known to occur on St.
Paul's Rocks, and they are certainly not a coral-reef. This absence of coral is not
easy to understand. When we were there (28.xii.37) the sea-temperature was
about 26°C, amply warm enough for reef-building corals, and as the rocks are less
than a degree north of the equator it is not be expected that there is a large range
of temperature. The rocks are set in the west-going Equatorial Current, which was

126

THE "ROSAURA" EXPEDITION

W
u

co u

Ph (DC

CO rrt

co £i

503

Cj CD

a s

bo C

CO i^

3 »

6 a ^

W -

^ .3

£3 <U

O
3 .

CC? -—.

CD
CD

T3
;-.
O

•O

+»

C(J

IS
£.5

CO >*

o g

EJ3

ffl

13
d

3
O

3
cd

525

CC}

0£

< 3

a
a
o

o
o

d .j

<u to
£ o

a a

o o

\r\ (M

fa?

6 "A

CO M

a ge

H
en

H
<J
H

CD +J

bo cd

-d d
^?d

(N CO

.d d

-t-> CD

p^ co

I 5

CD -+->

-t->

'35

co
aJ
bo

ho o>

T3 d

cd £

HH ■— 1

coCO

ft. dredge
Lore collect
rocks
d long line

o ^3

+» O
^ CD

3
3

H
-H

CD CD

3
—
o

'o
o

3
3
1

H
■0

H VN

NCfi (j

o
,d

5
.d

I I

^ ^^

,3
o

.2 ^

'.+_> CO

o o

m r

^ «

P-t £

^iz;

CO co

*CO

vo vO

°o

+»

rrt

^^ ^ ^ ^ ^ ^ £

co rt-

^4-

CO
M

o
u-j

o o

o
vO

o
iO

iO
iO

o
iO

o
vO

1>-

ti&

!zi

vb
o

M M

V cO

v o

V co
O

o
O
vO

O O

O O
vO vO

o
00

o
iO

ON
■4-

vO
CO

x :

3
d ^
"f o

THE "ROSAURA" EXPEDITION

127

'3.3

p 2

o o

in -p

o$ o

O T>
^g

£ :

ccj

Vh O

<L>
>

§-3

J O

£co o

.t3 CO

42 *

co o p>

en hh

4) N

to 6

£ ■s

•

• •

•

CI3

co n

O
O

to ■

1 1

fcuo

v©

M -*

-v

4->

.2 ©

-p 0)

13
O

a
1

1-)

+2 .

bo
T3

■H be

D bO

S/J

.8*

^ii

(75

co
1-1

O
CO

•

CO <s

•

. .

^' ^^' £>' ^*

^ *

^' ^>" bs ^' ^* $

in >o

CO
iO

IO w

CO M

O O
Tt-«0

O .

00 00

CO ro

00 00

vO
CO

O O

vO CO

00 00

°L

00 00

. -

» «

• «

•>

-. «

•

■

fcfc

00 00

"lOO

«o

C30

°~

°o

O O

t^ t^

t^.

CO CO

x :

K X

vO*

O M

'4-

»o

N N

co co

X X

'x :

t^oo

CN N

128

THE "ROSAURA" EXPEDITION

■*->

4->

o
o*
O

cc5

h- 1

1— t

in
o

a5
o

B
^

B
O
Q

o
o

o
_c

o ^

-M

tti

«B

6 £

c x.

<u cJ

>£

«Xi

> ^3

a3 O

.. .

° rtxl

outh
zue]
she)

xj Jj

Bfl

a}
H

-t->

C/3

6 o

6 S

O O

o v5

* *

Be * £ £

1
1

CO
00

o
CO

"co

o
O
00

-"f

vO
O

o
O

J b ^' £

-1 5
W >

S rt

o S 00

oft r

00
CO

■«-

(A
«o

«o

CO CO

•a

•s

"x'x

-<f

t>-

00 <J\

*i On
73 W

< «

■*■

»o

THE "ROSAURA" EXPEDITION

129

to ^

2
1§

-SS

<u -t->

«5 «J

(H TO

73 .
w

!zj

tf££

^^J

o3 \ »

Ph-°

en O

•-J 0) OJ

0) <T

. O

rQ ^ ^

03 a>

1/5 -3

to£

+»

T3 In

'%
O

Nianimaru, Gam

a. Salinity samp

b. Salinity samp

3 'Sh

T3 oS
fl CO
<D

CD T3

* s

to S

s'S
19

a .

?
e

m 0>

<U 4->

bog

■s!

CO CO

4J <D

CD bo
CO <N

O
C

5 <*<*

£ £

c^ r?

00 00
CO CO

00
N

N CO

•

< CQ

'■+3

t^.

i 3 o THE "ROSAURA" EXPEDITION

running at about 2 knots on the occasion of our visit. The islands are white with
guano, as sooty and noddy terns and boobies all nest there. I believe the only
plant life consists of unicellular algae, but a few insects and spiders have been recor-
ded, probably from round the birds' nests which are made of the green alga Caulerpa.
We were unable to land owing to the surf, but we got some plankton close to the rocks
and also dredged in about 40 fathoms from " Rosaura." This was not easy as the
bottom is excessively uneven, and also by the time that the dredge has been lowered
the ship has nearly drifted off into deep water. All we got was a sponge with numer-
ous epiphytes and commensal animals, and several fragments of algae, notably of
Caulerpa and Halimeda, genera which are both characteristic of warm seas and
especially of coral-reefs.

Our last haul was made with the Agassiz trawl in more than 700 fathoms between
the Canaries and the African mainland. Our position coincided with that of Station
41 in the 1910 cruise of the " Michael Sars " (Murray & Hjort, 1912), and the two
catches appear to have been very similar. We brought up many fish such as the
macrurid Coelorhynchus labiatus, and a large number of eels, Synaphobranchus pin-
natus, which were very much alive when they came up, about a score of them swim-
ming out of the net when it reached the surface ; this was after they had been
raised from a depth of 700 fathoms or so in just over half an hour. The catch inclu-
ded many abyssal Crustacea, and also a number of soft flexible sea-urchins of the
species Sperosoma grimaldii and Calveriosoma hystrix.

The whole collection from the cruise was deposited in the British Museum
(Natural History).

I owe a great debt of gratitude to the late Lord Moyne for taking me on this
cruise and for making these collections possible. I would also like to express my
sincere thanks for the friendly co-operation of the late Captain H. M. S. Laidlaw,
R.N.R., and the officers and crew of the " Rosaura," to whose collective skill the
successful completion of these hauls was due.

REFERENCES

Bruun, A. F. 1943. The biology of Spirula spirula (L.). Dana Rep. 24, 1943, pp. 1-44;

2 pi. and 13 text-figures.
Chun, C. 1910, 19*5- Cephalopoda. Deutsches Tiefsee Expedition, 18, Nos. 1 and 2.
Monro, C. C. A. 1939. Polychaeta of the Rosaura Expedition. Novitates Zoologicae, XLI.
Moyne, Lord. 1938. Atlantic Circle. Blackie.

Murray, Sir John, & Hjort, Johan. 1912. The Depths of the Ocean. Macmillan.
Parker, H. W. 1940. Undescribed anatomical structures and new species of reptiles and

amphibians. Ann. and Mag. Nat. Hist., ser. ii, 5 : 257-274.
Pickford, Grace E. 1939- On Melanoteuthis beebei Robson and the Vampyromorpha of

the M.Y. " Rosaura " collections. Ibid., ser. ii, 4 : 338, September, 1939.
Schmidt, J. 1922. Live specimens of Spirula. Nature, No. 2771, 110, 9.XU.22.

2. UNDER-WATER ILLUMINATION AND
ECOLOGY IN TROPICAL ESTUARIES

By JOHN S. COLMAN and L. H. N. COOPER

Little accurate information is available on the penetration of daylight in the
estuaries of tropical rivers. During the cruise of M.Y. " Rosaura " in the winter
of 1937-38, surface water samples were collected and brought back to Great Britain
for salinity analyses by the Government Chemist. These in some cases proved a
matter of difficulty owing to the excessive colour of the water. Since this colour
appeared likely to have considerable ecological importance, the Government Chemist
kindly returned the water samples to the Plymouth Laboratory where extinction
coefficients were determined in the Pulfrich Photometer with nine spectral filters
as described by Cooper & Milne (1938, 1939). The extinction coefficient, fi, is
defined by the equation

/id = 2-303 log 10 7 3«

where d is the length of the column of water measured in metres and I and I x
are the intensities of light of a given wave-length entering and leaving the water
sample. Three months elapsed between collection and the light measurements on
22nd March, 1938, so that there was time for bacterial action considerably to affect
the colouring matter. Admixture of salt water may lead to " clumping " or coagula-
tion of colloidal matter. 1 The results are sufficient to show that exceptional opacity
existed in the Barima river and in the Demerara river at Georgetown, both in
British Guiana (Tables I and II). Blue and green light were even more heavily
absorbed than were red and yellow. Data for Neal Point, the most turbid station
investigated in the Hamoaze (Tamar Estuary) near Plymouth, England, are in-
cluded for comparison. If the zone of active photosynthesis be considered as lying

1 In 1945 extinction coefficients of sea water determined by one of us with the Pulfrich photometer
were compared with coefficients determined on " brightness profiles " by an aerial photographic method
(Moore, i947» Table 8). In some cases determinations in the photometer were twice as great. Since
no firm explanation was apparent publication has been deferred. Poole & Atkins (unpublished work)
have since found a similar discrepancy between extinction coefficients determined with the Pulfrich
photometer and those measured in the sea with submarine photometers. It has become clear that the
extinction coefficient, measured on a beam of light passing through water enclosed in a tube in a Pulfrich
photometer, or probably in any similar visual or photoelectric laboratory absorptiometer, is a very
different physical quantity from the vertical extinction coefficient measured by a submarine photometer
exposed in the sea or the extinction coefficient determined photographically from brightness profiles.
Poole & Atkins explain the discrepancies as due to the forward scattering which they find light to undergo
in natural waters. Results with the Pulfrich photometer or any other laboratory absorptiometer
employing a beam of light in a tube have value, but they are not the same as those obtained by direct
observation at sea or from the air. (Added in proof, 23.ix.53.) There is yet another factor, in addition
to the effect of forward scattering from natural waters confined in glass tubes, which Cooper now accepts
without reserve. It has been found, following a suggestion of Dr. H. H. Poole, that the geometry of
the Pulfrich measuring assembly, as we have used it, makes for an apparent extinction coefficient larger
than that obtained by other methods. Comparison with results obtained by Cooper and Milne with the
same apparatus is legitimate. Comparison with any other procedure may not be.

ZOOL. 2, 6. 8

132

THE "ROSAURA" EXPEDITION

tation
No.

3 8b

39
40

47 a
47B

Name

Morowhana, Barima
River, British
Guiana

Junction of Barima
& Arouka Rivers,
British Guiana

Demerara River,
Georgetown,
British Guiana

Nianimaru, Gambia
River

Tendeba, Gambia
River

Table I.—Hydrographical Data

Date Position

Q.xii.37 . 8°i 5 'N., 59°45'W.

10.xii.37 • 2 m il es above (south
of) Station 3 8b

13.xii.37 .

2.L38 . 13 42'N., i4°58' W. ;

120 miles from mouth
3 .i. 3 8 . i3°42'N.,i 4 58'W. ;

120 miles from mouth
3.L38 . i 3 °27'N., i 5 °47'W. ;

50 miles from mouth

Surface

Depth

of
river
(m.)
12-20 .

State
of
tide •

Ebb ,

Salinity

(°/oo)

10-90

A. .

Tempera-
ture
(°C.)

. 8.3*

26°

7 ■

Flood ,

13-9*

••

Ebb .

o-oo

24°

o-oo

1 hr. .
after
high
water

2-63

24°

* Salinity is only approximate owing to uncertainty of end-point.

Station Spectral filter number
No. Centre of gravity of filter (mju.)

38B . Morowhana, Barima River

39 . Junction of Barima and

Arouka Rivers

40 . Demerara River, Georgetown
47B . Nianimaru, Gambia River

Neal Point Buoy, River
Tamar, England, low water

springs 16.ii.37
Ditto, 23.vi.37

Neal Point Buoy, River
Tamar, England, high water

neaps, 17.vi.37

Table II

s 43 S47
434 463

S50
494

S53
53o

S57
572

S61
619

S66-6
666

S72
729

S75
750

nine spectral 1
30-7 24-4
39-9 30-3

filters :
19-7
24-8

15-6
20 • 1

i3-i
16-9

10-9
i4'3

9'4

12-0

9-2
n-9

io-o
ii'i

27-2 24-6
6-5 5*9

23-1
5'25

21-4

4.85

19-5
4-80

19-3
4'45

18.5
4- 20

18-8
5'35

23.9
61

8 16-5 15-0 14-6 13-4 12-8

6-i8

2- II

5-48
I-8 7

5-01
1-71

4-62
i-58

4'63
i- 60

4*53
i-79

135 12-7

5-61 6-29
3-i8 3-86

>epths in metres at which intensity of light is reduced to 1 % of that incident on surface (no correction for surface
loss)

38B

Morowhana, Barima River

0-15

0-I9

0-23

0-29

035

0-42

0-49

0-50

0-46

Junction of Barima and
Arouka Rivers

0-I2

0'15

o- 19

0-23

0-27

0-32

0-39

0-42

Demerara River, Georgetown

CI7

0-I9

0-20

0'22

0-24

0-25

0-24

0-19

47B

. Nianimaru, Gambia River
Neal Point Buoy, River

0-71

0-78

o-88

o-95

0-96

1*04

I* 10

o-86

0*76

Tamar, England, 16.ii.37

0-24

0'27

0-30

o-33

o-35

—

o-33

o-35

Ditto, 23.vi.37

0-72

o-8i

0-89

0.96

0-96

0-98

—

0-79

0-71

,, 17.vi.37

2- IO

2-38

2" 60

2-8l

2-77

2-48

—

1-39

115

THE "ROSAURA" EXPEDITION

133

mainly above the level of 1% illumination (Table III) it must be confined to a
layer no more than 40 cm. thick.

Cooper & Milne (1939) found a close relation between the daylight factor and
zonation of algae on buoys moored in Plymouth Sound and the Hamoaze. The
daylight factors at the boundaries between the algal Zones I and II and between
Zones II and III lay at about 80% and 60% respectively. The depths in the
tropical estuaries corresponding to these daylight factors are set out in Table IV
in which the corresponding Plymouth data are given for comparison. The
same distribution of algae is not to be expected in such widely different environ-
ments, but even so the daylight factor shows very clearly how greatly the zone of
active photosynthetis must be compressed.

The maximum intensity of tropical sunlight is only about 30% greater than that
of the brightest summer day in Southern England, but there are two periods with
zenithal sunlight each year and the average solar altitude is considerably greater
(Atkins, Ball and Poole, 1937). Even so a dearth of phytoplankton and of the
animals dependent thereupon might be expected. A comparison with the condi-
tions actually found is thus of interest.

a. BRITISH GUIANA

Station 38 : Morowhana, Barima River, 8-9.xii.37

The river is tidal, about 300 metres wide and from 12 to 20 metres deep, and the
salinity on the ebb was io*90% . The water is apparently sometimes fresher than
this, for the local inhabitants are said to drink it. A haul with the 2-foot dredge
brought up mud and rotting vegetation, and nothing living except some barnacles
(young Balanus sp. in the Tetraclita stage) on a stone. The zooplankton, sampled
with a silk townet (56 meshes to the linear inch), was quite abundant and included
the following forms i 1

Pisces : Engraulidae : Larvae under 20 mm. . . . . 15

Sciaenidae :,,,,,,.... 22

Gobiidae : ,,,,„.... 1

Chaetodipteridae : Chaetodipterus faber (Brouss.) 14 mm. 1

Tetrodontidae : Colomesus psittacus (Schn.) 15 mm. . 2

Mollusca : Gastropoda, young, ...... Several

Crustacea : Decapoda : Sergestidae,

Penaeus sp., young .... Many

Zoeas,
Other larvae,

Amphipoda : in tubes, Several

Isopoda :........ 2

Copepoda. : including Labidocera fluviatilis Dahl "1

Acartia lilljeborgi Giesbr. f Many
Acartia giesbr echti Dahl. J

Chaetognatha : Sagitta sp., small ...... Many

Coelenterata : Scyphomedusae : Stomolophus meleagris, young, . 3

1 Such numbers of animals as are given in this paper must be regarded only as indications of relative
abundance.

134

THE "ROSAURA" EXPEDITION

The bulk of the catch consisted of zoeas, copepods and Sagittae. The presence
of the first two of these groups suggests that some planktonic plant-food must
have been present in spite of the low level of under-water illumination.

Station 39 : At the junction of the Barima and Arouka Rivers,
2 miles above Station 38, 10.xii.37

The salinity on the ebb was 8*3% . The river was remarkably full of fish, many
of which must have weighed several pounds. The plankton was here sampled
with a 2-m. stramin net (about 16 meshes to the linear inch), streamed from the
ship while at anchor. The catch was large but contained no zoeas, copepods or
Sagittae ; these would all have passed through the meshes of the net. The following
forms were caught :

Pisces : Engraulidae :......

Ariidae : T achy sums nuchalis (Gunther,) 30 mm.
Sciaenidae : Stellifer stellifer (Bloch), 10-35 mm.
Gobiidae : post-larval, 27 mm. .
Soleidae : Achirus sp., 23 mm. .
Cynoglossidae : Symphttrus sp., 33 mm.
Mugilidae : Mugil sp., 40 mm. .
Tetrodontidae : Colomesus psittacus (Schn.), 20 mm.
Also several larval fish, unidentified.
Crustacea : Decapoda : Xiphopenaeus kroyeri (Heller), small

Penaeus sp., very small
Sergestidae, young
Palaemon sp. .
Prawns, not determinable .
Amphipoda : Pseudoceradocus lutzi Schoemaker
Isopoda : .
Coelenterata : Scyphomedusae : Stomolophus meleagris, adults

>8o

1
1
1
1
>ioo

Many

Several

1
5
5

Station 40 : Demerara River at Georgetown, B.G., 13.xii.37

The river is here about a mile wide and the ship lay only one mile from the open,
though very shallow, sea. (There is a depth of only about 8 metres 15 kilometres
off shore). The tide ebbs and flows at about 5-6 knots, and on the ebb the salinity
was I3«9% . The water was so turbid that if one's hand was immersed to the
wrist the fingers were invisible. The tidal current in 7 metres of water no doubt
keeps the water well stirred right down to the bottom. This is suggested also if
we compare the extinction coefficients in the Demerara river with those in the
neighbouring Barima. Whereas the Barima is strongly coloured brown and shows
much the greater absorption in the blue and violet, the Demerara shows a more

THE "ROSAURA" EXPEDITION 135

uniform absorption through the spectrum, suggesting that a non-selective absorption
due to suspended detritus is super-imposed on selective absorption due to brown
pigments in true or colloidal solution in the water. (The tidal current in the Barima
is less than half that in the Demerara.) The very rapid extinction of under-water
illumination must confine active photosynthesis to a very shallow layer. Jenkin
(1937) found that the compensation point for Coscinodiscus excentricus lay at about
360 lux. This is a mean figure for 24 hours. For the hours of daylight it would
be considerably higher. If we take 140 kilolux as a round figure for midday tropical
daylight (Atkins, Ball & Poole, 1937) the daylight factor for Coscinodiscus would

exceed-^- =0-26%. The value 1% for the daylight factor over 24 hours is

therefore about as near as we can get at present. Reflection losses at the surface
may also reduce the incident light by 4 to 15% according to the degree of roughness.
Moreover land plants, and no doubt marine algal also, differ greatly in their response
to light. Even so, under the most favourable conditions the depth of the layer of
active photosynthesis could not have exceeded one- third of a metre, but nevertheless,
the river was full of algae in globular masses up to about an inch in diameter and
occurring every few inches ; they were covered with brown detritus but appeared
to be healthy, and were certainly not confined to the surface.

The plankton was sampled with the silk net, which soon became completely
clogged with these algae so that the catch was small, though it was probably fairly
representative qualitatively. It contained :

Pisces : Engraulidae ........ 2

Unidentified ........ 1

Crustacea : Decapoda : Penaeus sp., young .... A few

Cumacea ........ 2

Amphipoda : small I 1

Copepoda : including Acartia lilljeborgi Giesbr. . Many
Chaetognatha : Sagitta sp., small

Coelenterata : Scyphomedusae : Stomolophus meleagris, young . 3

Ctenophora ....... 1

Here again the presence of copepods seems to indicate unicellular plants on which
they feed, this in spite of the great opacity of the water. The two cumaceans
must represent a bottom fauna living in perpetual darkness. The occurrence of a
ctenophore in water with a salinity as low as I3'9% is perhaps worth noting.

The abundance of living organisms in these Guiana estuaries is much greater
than would be anticipated from the conditions of illumination, and a quantitative
ecological study should prove of considerable interest. Many marine animals
grow to a large size in the ocean abyss far below the region of active photosynthesis.
They depend on the rain of food from the illuminated waters above. Similarly
many of the animals described above from tropical rivers must have sought their
food in the vegetable and animal detritus brought down from the rain forest and
from more transparent inland waters.

136 THE "ROSAURA" EXPEDITION

b. GAMBIA RIVER, WEST AFRICA

Station 47 : Nianimaru, 2.L38

Nianimaru, where the river is nearly a kilometre wide, lies about 200 km. from
the mouth of the Gambia river, which remains tidal for yet another 210 km. Even
so the surface water was perfectly fresh, with a salinity of o-oo% just after high-
water. The plankton, sampled with the silk net on the ebb, was of a fresh-water
type, containing many reproducing cladocera, Bosmina coregoni, but nothing else.
The 2-foot dredge, shot in 9 metres, came up full of mud devoid of anything living.
The spectral absorption of the water was uniform and in no way remarkable. The
depth of 1% illumination lay at 94 cm. so that the conditions for photosynthesis
were much more favourable than in the Guiana rivers investigated, where the
corresponding depths ranged from 22 to 40 cm. (see Table III).

Table III. — Depth in metres at which total visible radiation was reduced to
1% of that incident on the surface. No correction for surface loss

Station

Depth

No.

Place

38B

Morowhana, Barima River

0-40

Jet. of Barima and Arouka Rivers

0-31

Demerara River, Georgetown

0-22

47B

Nianimaru, Gambia River

0-94

Neal Point Buoy

0-3-2-2*

* The lower figure applies only for short periods near low water during wet weather in winter. For
much of the year at this turbid station the depth of i % illumination is likely to exceed i • 5 m.

Table IV. — Depth in metres at which daylight factors of 80% and 60% would be
found (corresponding to the zonal boundaries of Cooper & Milne (1939, revision
of Table V, p. 3 9 5)

Zonal boundary .
Daylight factor .

•

I-II

80%

II-III
60%

Guiana rivers :

Morowhana, Barima R. .

0-0155

°'035

Jet. of Barima and Arouka

o-oii

0-027

Demerara, Georgetown .

•

o-oio

0-024

Gambia River, Nianimaru

•

• 040

O-IOO

English estuary :

Plymouth Sound .
Hamoaze Buoys Nos. 4, 7 &

o- 180
0-08-0- 10

0-356

o- 18-0-25

Station 48 : Tendeba, 3.1.38.

Tendeba is some 120 km. below Nianimaru and about 80 km. from the mouth
of the river. The surface salinity was 2-63% one hour after high-water. The
plankton was sampled with the stramin net and appeared to be not markedly

THE "ROSAURA" EXPEDITION

137

more numerous than and definitely not so varied as in the much darker waters of
the Barima river. The catch contained :

Pisces : Clupeidae : Pellonula afzeliusi Johnels

Ilisha africana (Bloch) 1 15-130 mm.
Sciaenidae : Johnius elongatus (Bowdich) 58-65 mm.
Atherinidae : Larvae, <I5 mm.
Crustacea : Decapoda : Callinectes latimanus Rathbun, 1 female

Penaeus sp., immature
Crangonidae, immature
Mysidacea ....

Isopoda .....
Nematoda ......

>200

>25

I
Many

> >
Abundant

1
1

1 Several other crabs, presumably of this species and up to 10 cm. across, were seen at the surface
drifting with the tide.

No copepods or Sagittae were caught, but they may have passed through the meshes
of the net. The plankton at this station was remarkably colourless.

We are indebted for the names of species to the following specialists : Dr. I.
Gordon (Decapods and Amphipods), Dr. J. Harding (Copepods), Lieut.-Col. W. P. C.
Tenison, D.S.O., and Mr. D. W. Tucker, (Fish) and Capt. A. K. Totton (Coelenterates) ;
also to Dr. W. R. G. Atkins, F. R. S., for discussion of the conditions of illumination
and photosynthesis.

SUMMARY

Waters from the tidal estuaries of the Rivers Barima and Demerara in British
Guiana show extremely heavy absorption of light of all wave-lengths ; never-
theless, in spite of the unfavourable conditions for photosynthesis, the zooplankton
is both varied and abundant. These observations in rivers of British Guiana are
compared with similar ones made in the Gambia River, W 7 est Africa.

REFERENCES

Atkins, W. R. G., Ball, N. G., & Poole, H. H. 1937. The photo-electric measurement
of the diurnal variations in daylight in temperate and tropical regions. Proc. Roy. Soc,
ser. A, 160, No. 903, pp. 526-539.

Cooper, L. H. N., & Milne, A. 1938. The ecology of the Tamar Estuary. II. Under-
water illumination. Journ. Mar. Biol. Assoc. 22 : 509-528.

1939- The ecology of the Tamar Estuary. V. Under- water illumination. Revi-
sion of data for red light. Ibid. 23 : 391-396.

Jenkin, P. M. 1937. Oxygen production by the diatom Coscinodiscus excentricus Ehr. in
relation to submarine illumination in the English Channel. Ibid. 22 : 301-343.

Moore, J. G. 1947. The determination of the depths and extinction coefficients of shallow
water by air photography using colour filters. Phil. Trans. Roy. Soc, ser. A, 240 : 163-
217.

3. THE ECHINODERMATA

By D. DILWYN JOHN and AILSA M. CLARK

With Plate 6 and Text-figs. 1-12.

The collection of Echinoderms taken by the *' Rosaura " includes specimens from
off Greenland and north-east America, as well as from the West Indies and one
station off the north-west coast of Africa.

Some of the Echinoids were identified by the late Dr. Th. Mortensen before the
war. The work was continued by one of us (D. D. J.), who described those species
which are new and commented on various other species ; the remarks about the
new species and the remaining specimens, the preparation of the figures and the
final compilation were done by the other (A. M. C).

The species taken are as follows (those marked with an asterisk being mentioned
in the text) :

Name
Asteroidea
Ctenodiscus cnspatus (Retzius)

Astropecten marginatus Gray
Astropecten antillensis Liitken
Astropecten articulatus (Say)
*Tethy aster vestitus (Say)
*Luidia rosaurae sp. n. .
*Luidia barimae sp. n.
Luidia senegalensis (Lamarck)
Luidia clathrata (Say)

Anthenoides piercei Perrier
Nymphaster arenatus (Perrier)

Goniaster cuspidatus Gray .

Station

*Ceramaster granulans forma
(Sladen)
Echinaster sentus (Say)
Henricia sanguinolenta (O. F. Miiller) .
Crossaster papposus (Linnaeus) .

*Brisingella coronata (G. O. Sars) .
♦Brisingidae — fragment

Zoroaster ackleyi Perrier
* Stephanasterias albula (Stimpson)

Leptasterias polaris (Miiller & Troschel)

Number

of
specimens

11 (Newfoundland) and 6 (S.W. Green- 54

land)

36 (British Guiana) .... 3

35 (off R. Orinoco) .... 2

35 ,. .... 1

35 „ .... 1

35 „ .... 3

35 „ .... 2

36 (British Guiana) .... 1
22 (British Honduras) and 37 (Brit. 2

Guiana)

35 (off R. Orinoco) and 36 (Brit. Guiana) 4

26 (Brit. Honduras) and 49 (off N.W. 3

Africa)

26 (Brit. Honduras) and 35 (off R. 2

Orinoco)

balteatus 49 (off N.W. Africa) .... 9

35 (off R. Orinoco)

10 (South Labrador)

10 ,, ,, and

land)
49 (off N.W. Africa)
26 (British Honduras)
26
6 (S.W. Greenland)
9 (South Labrador)

1 1 (Newf ound-

1
1
2

1 arm

1
2

140

THE "ROSAURA" EXPEDITION

Name

Ophiuroidea

Astrophyton muricatum (Lamarck)

Astroporpa annulata Lutken

Ophiacantha pentacrinus Lutken .

Ophiacantha metallacta H. L. Clark

Ophiacantha bidentata (Retzius) .

Ophiomitra valida Lyman .
*Ophioplinthaca grenadensis sp. n.

Ophioplinthaca incisa (Lyman) .

Amphiura sundevalli (M. & Tr.) .
*Arnphipholis gracillima (Stimpson)

Ophiactis savignyi (M. & Tr.)
*Ophiactis abyssicola (Sars) .

Ophiopholis aculeata (Linnaeus) .
*Ophionereis dolabriformis sp. n.

Ophiocoma echinata (Lamarck) .

Ophioderma brevicaudum Lutken
*Amphiophiura metabula H. L. Clark

Ophiocten sericeum (Forbes)

Stegophiura nodosa (Lutken)
*Ophiomusium validum Ljungman

Ophiolepis elegans Lutken .

Ophiura sarsi Lutken
*Ophiernus adspersum Lyman

Number

Station

specimens

. 35 (off R. Orinoco)

• 35 » ...

34 (Grenada) ....

• 34 ....

116

11 (Newfoundland)

34 (Grenada) ....

34 ....

. 34 „ and 26 (Brit. Honduras)

1 (Greenland) ....

22 (British Honduras) .

. 31 (Mosquito Bank, W.I.)

. 49 (off N.W. Africa)

1, 4, 6 (Greenland) and 10 (S. Labrador

) c. 34

. 35 (off R. Orinoco)

16 (Grand Cayman Id.)

37 (British Guiana)

. 34 (Grenada) and 26 (Brit. Honduras)

1 (Greenland) ....

10 (S. Labrador) ....

26 (British Honduras) .

. 31 (Mosquito Bank, W.I.)

6 (Greenland) and 11 (Newfoundland)

. 34 (Grenada) and 26 (British Honduras)

Echinoidea

Tretocidaris bartletti (A. Agassiz)
Salenocidaris rarispina A. Ag.

Plesiodiadema antillarum (A. Ag.)

Sperosoma grimaldii Koehler
Calveriosoma hystrix (W. Thomson)
Phormosoma placenta W. Thomson)
Tripneustes ventricosus (Lamarck)
Strongylocentrotus drobachiensis (O.
Miiller)

Holothuroidea

Holothuria mexicana Ludwig
Mesothuria gargantua Deichmann
Phyllophorus pellucidus (Fleming)
Chiridota laevis (Fabricius)

35 (off R. Orinoco)

26 (Brit. Honduras) and 49

(off

N.W.

Africa)

26 (Brit. Honduras) and 49

(off

N.W.

Africa)

49 (off N.W. Africa)

49 „ „ . .

49 . .

18 (Honduras)

1, 5 (Greenland), 9, 10 (South Labrador)

c. 30

and 12 (Nova Scotia)

16 (Grand Cayman Id.)

26 (British Honduras) .

6 (Greenland)

5 and 6 (Greenland) .

Crinoidea
Tropiometra carinata (Lamarck)

. 37 (British Guiana)

THE "ROSAURA" EXPEDITION 141

Class ASTEROIDEA
Family Astropectinidae

Tethyaster vestitus (Say)

Asterias vestita Say, 1825 : 143.
Sideriaster ? vestitus, Verrill, 191 5 : 193.

St. 35. 9° 25' N., 59 52' W. (off the mouth of the river Orinoco) ; otter trawl ;
86 m. One specimen.

R = 70 mm., r ■= 16 mm. R/r = 4-3/1.

This name was eventually arrived at after a great deal of correspondence with
Mr. Austin Clark of Washington. He has recently received a huge Astropectinid
from off North Carolina which answers very well to Say's description of Asterias
vestita as far as that goes. This specimen has R = 250 mm., with a diameter of
about 18 inches, probably as, compared with one of 14 inches for Say's lost type from
Cape May, New Jersey, and it almost certainly represents the same species. A
notable feature of this form is the development of large spatular ventro-lateral and
infero-marginal spines, similar to, but less well-developed than, those of Ar chaster
magnificus Bell from St. Helena in the south-east Atlantic. One of Bell's types
with R = 215 mm., has these spines so large as to overlap, producing a very " shaggy"
appearance, each spine being about 6 mm. long, whereas in the even larger specimen
from North Carolina these spines are only about 4 mm. long. There seems to be no
other significant difference between the two forms. Caso (1947 : 225) appears to
think that magnificus is synonymous with Sideriaster grandis Verrill from the Gulf
of Mexico. This opinion was shared by Mr. Clark and Dr. W. K. Fisher, but on the
evidence of a number of specimens from the type locality Mr. Clark now concludes that
grandis also is distinct from both magnificus and vestitus, being differentiated by the
pointed (rather than truncated) form of the ventro-lateral and infero-marginal spines,
developed only at a late stage (when R is more than 150 mm.) as in the latter. This
is coupled with other features such as the relatively shorter dorsal paxillae in grandis.

The specimen taken by the " Rosaura " is too young for the ventro-lateral spines
to have developed, so that it is not absolutely certain whether it represents vestitus
or grandis. However, Mr. Clark has a large specimen with R = 160 mm. from
Porto Rico which has small blunt ventro-lateral spines like vestitus. He therefore
believes that vestitus occurs in the West Indies as well as up the East coast of the
United States as far as New Jersey while grandis is limited to the Gulf of Mexico.
This is similar to the distribution of Astropecten articulatus (Say), which extends from
North Carolina (? New Jersey) south to Uruguay with a subspecies valencienni in
the Gulf of Mexico, as pointed out by Mr. Clark. Whether the two forms of Tethy-
aster are specifically or subspecifically distinct is controversial.

As for the generic name used here, it was finally decided that Tethyaster Sladen
with genotype Asterias subinermis Philippi embraces all the species formerly inclu-
ded in Sideriaster Verrill and Monaster Sladen as well as Anthosticte Fisher. This

i 4 2 THE "ROSAURA" EXPEDITION

amalgamation will be discussed in more detail in a separate paper on Tethyaster
by A. H. Clark and A. M. Clark. It is sufficient here to list the species included
in the genus.

subinermis (Philippi) (genotype)
vestitus (Say)
grandis (Verrill) .
magnified (Bell) .
aulophora (Fisher)
canaliculatus (A. H. Clark)

Mediterranean, Bay of Biscay to Mauritania.

New Jersey to Venezuela.

St. Helena and Ascension.

Gulf of Mexico.

Philippines.

Gulf of California.

gigas (Caso) (possibly only the Gulf of California.

large form of canaliculatus)
pacei (Mortensen) . . . South Africa.

Family Luidiidae

Luidia rosaurae sp. n.

(PI. 6 fig. i ; Text-figs, i and 2)

St. 35. 9 25' N., 59 52' W. (off the mouth of the river Orinoco) ; otter trawl ;
86 m. Three specimens.

Diagnosis. A species of Luidia with five rays belonging to the A Item at a group
of Doderlein ; no enlarged dorsal spines present ; the paxillae of the two outermost
rows larger than those of the supero-marginal series ; pedicellariae present on many
of the dorsal paxillae particularly the lateral ones, as well as on the infero-marginal
plates ; four adambulacral spines on each plate, the two outermost arranged in a
line parallel to the furrow ; three- or four-valved pedicellariae on the ventro-lateral
plates of many segments, especially proximally ; one large marginal spine on each
infero-marginal plate with small appressed ones on the ventral surface of the plate.

Description. The three specimens are all rather broken. Each has five rays.
In the only one which has a complete and original (i.e., not regenerated) arm, R is
about 53 mm. and r is 6-5 mm. R/r = 8/1.

The three outer rows of paxillae on the arms are larger than the remainder ; the
paxillae of the second and third rows are larger than those of the first. The paxillae
of the first (i.e., supero-marginal) series, are rectangular, being slightly elongated.
Those in the proximal part of the arm consist of 25-30 spinules, five to eight of which
are in the centre, the remainder around the edge. In the centre of most there is
also a small bivalved pedicellaria. The paxillae of the second and third rows are
rectangular, being slightly longer in a transverse than a longitudinal direction ; those
shown in fig. 2 are foreshortened. At the base of the arm they consist of about 30
to 38 spinules, of which 14 to 16 are in the centre. The central spinules are heavier
and blunter than the peripheral ones. A small number of these paxillae have two-
valved pedicellariae among the central spinules.

The paxillae occupying the centre of the arm are smaller and roughly circular.
At the base of the arm there are eight or nine rows of them, of which those of the

THE "ROSAURA" EXPEDITION

*43

outer row" on either side are slightly larger than the remainder. They consist of
a central group of 14 to 16 spinules, of which those nearest the centre are heavy
and blunt, and a peripheral circle of about the same number of spinules. In the
paxillae in the middle of the arms the contrast between the central and peripheral
spinules is more marked. There are up to eight central spinules, of which some or
all are considerably heavier than the peripheral ones, of which there are about ten.
These paxillae are very variable. They do not form regular rows. Those at the
centre of the disc are of the same nature but smaller and more crowded.

The innermost adambulacral spine arises within the furrow. It is long and com-
pressed and strongly recurved. A second and similar spine, not quite so recurved,

Fig. 1. Luidia rosaurae sp. n. Type. Ventral view of one side of two adjacent joints,
that on the left having been denuded with sodium hypochlorite. (The spines are
displaced sideways to avoid foreshortening as much as possible ; the arrow points
towards the mouth.)

occurs just outside it. External to this are two slightly shorter spines standing side
by side ; they are straight and compressed and taper from a wide base to a narrow
tip. The outer part of the adambulacral plate is paitly overlapped by the ventro-
lateral plate which frequently carries a large and conspicuous three-or four-valved
pedicellaria as well as several spinules. The pedicellariae are most numerous near
the interbrachial angles.

The underside of each infero-marginal plate is raised into a strong ridge, the
adjacent ridges being separated by deep grooves. The outer surfaces of the ridges
are occupied by flattened spinelets of medium but varying lengths ; fine spinules
arise from the edges of the ridges. Near the outer edge there are often one or two
pedicellariae, smaller than those of the ventro-lateral plates, of two, or more rarely
three, valves. One strong, somewhat flattened, tapering spine nearly 2 mm. long

i 4 4

THE "ROSAURA" EXPEDITION

arises from near the upper edge of each plate forming a regular border to the arm.
There is a paxilla-like group of spinules with one or two bi-valved pedicellariae above
each spine, at least at the base of the arm. The outer parts of the arms are rather
badly damaged.

When freshly preserved one of the specimens was straw coloured, with a brown
patch in the centre of the disc continuous with broad brown bars running a short
distance down the middle of each arm ; there were also brown bands across the
outer parts of the arms at intervals. The other two specimens were brown, but with
deeper colour in areas corresponding to the brown parts of the first specimen. (After
15 years all colour has been lost.)

Fig. 2. Dorsal view of the top of two infero-marginal plates (seen in profile because of distortion
in preservation causing an abrupt bend just above these plates) and the adjacent paxillae.

Remarks. This species is obviously most closely related to Luidia scotti Bell
(1917 : 8), which is described fully in the third note on Asteroids in the British
Museum by A. M. Clark (1953 : 383). The types of L. scotti were taken off Rio
de Janiero in 40 fathoms (73 metres). They differ from L. rosaurae chiefly in having
much thicker, more stumpy spines and spinules on the ventral plates, a second
enlarged marginal spine above the ambital one on the infero-marginals, and in
having fewer ventro-lateral and infero-marginal pedicellariae and no dorsal ones.

This last character — the abundance of pedicellariae — is probably not of sufficient
value to be used alone for specific distinction but the three points taken together are
enough, on the evidence of the material available, to distinguish the two forms as
separate species.

The second point is interesting since in the types of Luidia doello-juradoi Bernas-
coni (a synonym of L. scotti in the opinion of A.M. C.) from off Buenos Aires (Bernas-
coni, 1 941 and 1943) the two large marginal spines on each infero-marginal plate are
said to be nearly the same size, the upper being only slightly smaller. In the types

THE "ROSAURA" EXPEDITION 145

of L. scotti the upper spine averages only one-third of the length of the lower one,
although occasionally it is two-thirds as long.

Luidia armata Ludwig (1905 : 85) from the Gulf of Panama also has much in
common with Luidia rosaurae but perhaps even more with L. scotti. It has ventro-
lateral pedicellariae only in the interbrachial angles, but also has dorsal pedicellariae
on the supero-marginal paxillae. Figures of this and of the other related species
of Luidia described from Central and South America will probably show that they
cannot all be maintained as valid species. This may be true of Luidia armata
Ludwig, L. ludwigi Fisher (1906 : 122 and 1911 : 113), L. scotti Bell, L. doello-juradoi
Bernasconi and L. rosaurae, which could represent local forms of a single species
occurring on both sides of South and Central America. However, without Pacific
specimens for comparison this suspicion cannot be verified.

Luidia barbadensis Perrier (1881 : 29) also appears to be related to L. rosaurae.
It is apparently usually six-rayed, whereas the three specimens of L. rosaurae are
all five-rayed. The only pedicellariae described in L. barbadensis are said to be on
the adambulacral plates (probably these are actually on the ventro-lateral plates
as in related species) and in the interradial areas (Verrill, 1915 : 205-7). ^ n Luidia
rosaurae they also occur on the outer paxillae of the arms dorsally and on the outer
and lower surfaces of the infero-marginal plates. The infero-marginal plates of
L. barbadensis bear two large spines on the margin of the ray. In L. rosaurae there
is only one large marginal spine. There are also differences in the arrangement and
proportions of the adambulacral spines of the two species. In the type of L. barba-
densis with R = 125 mm. there are only two spines behind the furrow spine not
three as in L. rosaurae, although Verrill (1915 : 207) describes an " Albatross "
specimen of L. barbadensis with four adambulacral spines arranged as in this new
species.

A good series of specimens of Luidia barbadensis or detailed figures of the types
are needed to define more clearly the relationships of the species.

Perrier's other six-rayed species Luidia convexiuscula (1881 : 30), is based on a
specimen with R only 28 mm. There appears to be no significant difference by which
it can be distinguished from L. barbadensis, taking into account the very considerable
size discrepancy between the types of the two species. The presence of three- valved
pedicellariae only in the arm angles, only one enlarged marginal spine and the poor
development of the adambulacral spines would be expected in such a young speci-
men.

Luidia barimae sp. n.

(PL 6, fig. 2 ; Text-figs. 3 and 4)

St. 35- 9° 25' N., 59 52' W. (off the river Orinoco) ; otter trawl ; 86 m. Two
specimens.

Diagnosis. A ten-rayed species of Luidia belonging to the Quinaria group of
Doderlein, with all the dorsal paxillae irregularly arranged, some of them with one
or more pedicellariae ; no spino-paxillae ; three adambulacral spines in a row at

146

THE "ROSAURA" EXPEDITION

right angles to the furrow ; one to four ventro-lateral pedicellariae ; no furrow pedi-
cellariae except on the mouth plates ; infero-marginal plates extending on to the
dorsal side, with four (sometimes three) large spines which alternate in position
on adjacent plates.

Description. The two specimens are ten-rayed. Most of the arms are broken
off near the disc and lost. The longest arm remaining attached to a specimen is
65 mm. long ; its tip is regenerated so that its length cannot be taken as a measure-
ment of R. The radius of the disc is about 13 mm. The arms are constricted to
a width of 9-5 mm. where they join the disc ; farther out they are 13 mm. wide.

None of the paxillae of the arms are arranged in regular rows. Those near the
margin are somewhat bigger than, but otherwise similar to, those in the centre.

Fig. 3. Luidia barimae sp. n. Type. Ventral view of one side of two joints, that
on the left having been denuded with sodium hypochlorite. (The arrow points
towards the mouth.)

Each consists of a peripheral circle of eight to twelve widely spreading spinelets
and a central spinelet, a little stouter but no longer than the peripheral ones. One
of the peripheral spinelets is frequently replaced by a small bi-valved pedicellaria.
There are no spino-paxillae.

The paxillae in the centre of the disc are small and much more crowded than those
of the arms. In many the central spinelet is missing. Nearly every paxilla is sup-
plied with a pedicellaria, while many have two. The pedicellaria sometimes arises
from near the centre of the paxilla.

The madreporite is 2-5 mm. in diameter and may be trefoil-shaped ; it arises
from very near the edge of the disc. A few paxillae appear to arise from it, but in
reality project from between the lobes, not from the madreporite itself.

The bigger elements in the armature of the adambulacral, ventro-lateral and
infero-marginal plates form one continuous transverse line from the furrow to the
abactinal surface. The adambulacral armature consists of spines and pedicellariae.

THE "ROSAURA" EXPEDITION

147

There is a strongly curved furrow spine, about 3 mm. long ; it is compressed to a
blade-like form and its end is blunt. It is followed by another blunt compressed
spine which is broader and may be longer, up to 4-5 mm. long ; it may be bent at its
base but is thereafter straight. There is a third spine which is nearly as long but
much narrower, being only slightly compressed. A few spinelets usually arise from
near the base of this spine on the adoral side. There may be a fourth and smaller
spine borne on the inner end of the ventro-lateral plate, but it is usual for these
plates to carry two or three large bi-valved pedicellariae, of which the outermost,
about 1 mm. long, is larger than the inner ones. Small spinules may arise around
the base of the pedicellariae, especially, in the proximal part of the arm, where there
may even be four rather than three pedicellariae on each plate ; in the distal part

Fig. 4. Dorsolateral view of three infero-marginal plates and the adjacent paxillae.

of the arm there is only one. There are no pedicellariae in the furrow as are found
in L. avicularia and L. Integra.

The infero-marginal plates encroach strongly on the abactinal surfaces of the arms.
They are raised into strong ridges. In the proximal part of the arm there are four
large spines on each plate. Three are on the lateral edge, one above the other, the
lowermost shorter than the second and third which may be 6 mm. long. The
fourth and longest spine, up to 8 mm. long, arises from or near the inner edge of the
plate on the abactinal surface. The spines of adjacent plates arise at different
levels as a rule, so that the uppermost spine of one plate is based on the extreme
adradial edge, while that of the next plate is somewhat inset, the lower spines being
correspondingly displaced on every second plate. On each row of infero-marginal
plates there are thus eight series of spines, those of any one series all arising at the
same level and usually only occurring on alternate plates. On those plates where
the uppermost spine is not on the extreme inner edge of the plate there may be only
three spines in the distal part of the arm. Numerous spinules occur in a thick row
on the proximal and distal edges of the infero-marginal plates ; they point into
the furrows between the plates. One or two small bi-valved pedicellariae occur on

Zool. 2, 6. o

I4 8 THE "ROSAURA" EXPEDITION

most of the plates, often on the innermost abactinal edge when there is no spine
in that position.

Pedicellariae are numerous on the jaws and in the small interradial areas.

The colour of the spirit specimens is greyish-brown above, light below.

Remarks. This species appears to be most nearly related to Luidia heterozona
Fisher from West Africa (of which L. mortenseni Cadenat is a synonym), although
that species is included by Fisher (1940 : 268) in the Ciliaris group since its super-
marginal plates are markedly larger than the adjacent lateral paxillae.

Doderlein's genealogical tree (1920 : 223) indicates that the Ciliaris group is
very closely related to the subgenus Integraster of the Quinaria group (in which
L. barimae seems to belong), and indeed Luidia heterozona is as much, if not more,
in accord with the ten-rayed Quinaria group species such as L. integra Koehler,
L. avicularia Fisher and L. moroisoana Goto as with the other species of the Ciliaris
group, none of which has more than eight arms. In the subgenus Integraster the
supero-marginal paxillae are larger and fewer in number than the adjacent rows of
paxillae, even it not so conspicuously different as in L. heterozona. In L. barimae
the lateral paxillae are much more irregularly arranged.

The only other notable difference between L. heterozona and this new West Indian
species is in the number of infero-marginal spines, of which there are commonly
four on each plate in L. barimae but only two or three in L. heterozona, although the
conspicuous alternation in the position of these spines is similar. The number of
pedicellariae between the outermost adambulacral spine and the lowermost infero-
marginal one may be as many as three in some specimens of L. heterozona according
to Madsen (1950 : 204), although the number is usually one. On the whole it
seems better to regard Luidia heterozona as also belonging to the Quinaria group,
it being the east Atlantic counterpart of L. barimae. Whether the differences
between the two are less than specific, as with the east American Luidia alternata
and its subspecies numidica from West Africa, remains to be seen from a greater
range of West Atlantic material, showing the variation particularly of the armature of
the infero-marginal and ventro-lateral plates.

The Indo-Pacific species Luidia integra, L. avicularia and L. moroisoana are
easily distinguished from L. barimae by the possession of spino-paxillae, and in
having a long-jawed pedicellaria within the furrow on the adambulacral plates.

Family Goniasteridae

Ceramaster granulans forma balteatus (Sladen)

Pentagonaster balteatus Sladen, 1891 : 688, pi. 25, figs. 1-5.

Pentagonaster hystricis von Marenzeller, 1893 : 4, pi. 1, fig. 2, pi. 2, fig. 2 ; Ludwig, 1897 : 179,

pi. 8, fig. 2.
Pentagonaster gosselini Perrier, 1894 : 399, pi. 26, fig. 4 ; 1896 : 45 ; Koehler, 1909 : 84, pi. 1,

fig. 9.
Pentagonaster kergroheni Koehler, 1896 : 63, pi. 2, figs. 8-10.
Ceramaster balteatus, Mortensen, 1927 : 82, text-fig. 45.
? Pentagonaster haesitans Perrier, 1894 : 397, pi. 23, fig. 7, pi. 25, fig. 2.

THE "ROSAURA" EXPEDITION 149

St. 49. 28 25' N., 13 34' W. (between Fuerteventura Island and Africa) ; agassiz
trawl ; c. 1300 m. Nine small specimens.

The typical north-European form of Ceramaster granulans (Retzius), from depths
as shallow as 20 metres, is nearly pentagonal in outline, with the short triangular
rays projecting abruptly. The R/r ratio equals c. 1-3 to 1-5/1, rarely more. The
marginal plates of both series each have a bare patch, which is relatively larger
in the young sea-star. There are usually three or four short, thick furrow spines and
two, sometimes three, shorter spines in the second row backed by several rows of
granules. The variations have been studied in detail by Grieg (1907 : 22 — 32, text-
figs. 1 and 2). It has been reported from off Morocco by Perrier and from the east
coast of North America by Sladen and Verrill.

From south-west of Ireland to north-west Africa and the Mediterranean, at
depths of more than 1,000 metres, is found a relatively longer-armed form with R/r
averaging 1*7/1 and with the interbrachial arcs regularly curved, not angular. The
marginal plates are more or less completely covered with granules and the furrow
spines number from four to six on each adambulacral plate, with three spines in the
second series backed by granules.

Several names have been given to such specimens, of which Pentagonaster bal-
teatus Sladen has priority. Sladen's type was dredged in 750 fathoms (1,372 metres)
in 51 01 ' N., n° 50' W. (south-west of Ireland). It has all the marginals completely
covered with granules, R/r = 1*7/1, and there are six furrow spines on each plate,
the adoral one being inset and shorter than the others.

Perrier originally described Pentagonaster gosselini as having only three or four
furrow spines, but later (1896) mentions specimens with four to six adambulacral
spines. The supero-marginal plates may be covered with granules or have bare
patches according to Koehler (1909) ; Perrier says that their granulation is " fugace."

" Pentagonaster " kergroheni Koehler from the Bay of Biscay in 1,710 metres has
only the distal supero-marginals with a small bare patch free of granules. R/r = 17/
1 and there are five furrow spines. Ludwig (1897) declares that P. kergroheni is a
synonym of P. hystricis von Marenzeller from the Mediterranean, which may also
have small bare patches on the marginal plates. This seems very likely, although
the Mediterranean form apparently has the granulation more spaced out than in
Atlantic specimens. However, Ludwig is mistaken in thinking that P. concinnus
Sladen and P. greeni Bell might also be synonyms of P. hystricis, for the former
having only marginal granules on the dorsal plates is synonymous with Plinthaster
perrieri (Sladen) as other authors have pointed out, while P. greeni does not have
tabulate dorsal plates and like Ceramaster placenta (Muller & Troschel) might be
referred to Peltaster Verrill.

Ludwig also says that older specimens of P. hystricis tend to have more numerous
and larger bare patches on the marginals than young ones, which is the converse of
what occurs in typical granulans. However the " Rosaura" specimens, the largest
of which has R only 15 mm., all have bare patches on the supero-marginals and five
of them on all the infero-marginals also. The other four, including the largest
one, have the proximal infero-marginals completely covered with granules, and
only the last two or three have a small bare patch. The R/r ratio varies between

150 THE "ROSAURA" EXPEDITION

1*45 and 1*7/1, averaging i-6. The width of the marginals also seems to vary with
the length of the arms, as they are broader when the shape is more pentagonal.

As for North American specimens of Ceramaster granulans, those collected by the
" Challenger " south of Halifax, Nova Scotia, in 156 metres, differ little from the
north European form except in having relatively longer arms, so that R/r averages
1-7/1. " Pentagonaster " eximius Verrill (1895) is a synonym of Ceramaster granu-
lans, for Verrill's conception of the latter obviously approximates more to balteatus
than to the typical form, as can be seen from his comparison. It seems quite probable
that a longer-rayed form somewhat similar to balteatus occurs beside the typical
one on the east coast of North America, just as typical granulans is found off
North Africa according to Perrier, as well as the longer- armed form.

Family Brisingidae

St. 26. 17 53' N., Sy° 44' W. (off British Honduras) ; agassiz trawl ; c. 900m.
One damaged arm fragment consisting of 46 joints, 95 mm. long.

At one end of the fragment the first and second adambulacral ossicles have, lying
above and fused to them, two marginal plates, showing that this is the proximal
part of an arm. The first marginal plate on one side is the same size as that on the
other. The first two pairs of ambulacral plates are shorter than the remainder,
which are each about 2-25 mm. long. The soft tissue is stripped off the first seven
joints, but some remains on the dorsal side between the eighth and the sixteenth
joints. Five costae, consisting of light ossicles with no spines, occur in that interval,
that is, one to about every two joints. Beyond the sixteenth joint the arm consists
of ambulacral and adambulacral ossicles, spines and shreds of tissue.

Each of the adambulacral ossicles carried a moderate-sized spine on its lower
edge, a little nearer to the distal than the proximal end. Larger spines than these
arise from the distal end of the outer edge of the adambulacral ossicles beyond the
seventh ; they appear to occur, with slight irregularities, on alternate ossicles.
None is complete, the longest remaining being about 4-5 mm. long. The skin
coverings of most are torn off but they are covered with numerous small crossed
pedicellariae.

We cannot with confidence assign this fragment to any of the three species of the
Brisingidae known from the Caribbean region : Hymenodiscus agassizi Perrier,
Odinia antillensis A. H. Clark and Freyella mexicana A. H. Clark. What remains
of the adambulacral spines are unlike those described for 0. antillensis and F. mexi-
cana. They are not unlike those of H. agassizi, and the structure of the arm is
very like what Perrier describes for that species (1884 : 189, pis. 1 and 2) : the ambul-
acral ossicles have an exactly similar process for articulation with the adambulacral
ossicles. Fisher (1918 : 104, figs. 1 and 2) describes how the first marginals of H.
agassizi which meet in an interbrachium are unequal in size. The first marginals
of this single arm are of equal size ; it may be that each met in its interbrachium
a smaller first marginal. Fisher found that the abactinal integument of the rays of
H. agassizi contained holothuroid plates. There are none in this specimen ; there
are on the other hand a number of light costae.

THE "ROSAURA" EXPEDITION 151

Brisingella coronata (G. O Sars)
Brisinga coronata G. O. Sars, 1871:5; 1875:1.

St. 49. 28 ° 25' N., 13 34' W. (between Fuerteventura Island and Africa) ;
agassiz trawl ; c. 1300 m. One disc.

The disc is of a young specimen, for it is only 12 mm. in diameter and the spines
on its abactinal surface are fine and fairly far apart. There were eight arms.

Family Asteriidae

Stephanasterias albula (Stimpson)

Asteracanthion albulus Stimpson, 1853 : 16, pi. 1, fig. 5.
Stephanasterias albula Heding, 1935 : 38.

St. 6. 6o° 06' N., 45 25' W. (off S.W. Greenland) ; three-feet dredge ; no m.
Two specimens.

At first sight the larger specimen with R = n mm. appears to be five-rayed,
but one of the larger arms has a very short arm coalesced with its proximal portion.
The smaller specimen is six-rayed.

Class OPHIUROIDEA
Family Ophiacanthidae
Ophioplinthaca grenadensis sp. n.

(Text-figs. 5-7.)

St. 34. 12 05' N., 6i° 49' W. (of St. George, Grenada) ; agassiz trawl ; 720-800 m.
One specimen.

Diagnosis. A species of Ophioplinthaca with oval, medium-sized, bare, radial
shields ; long, tapering, rugous disc-spines not concealing the scales (at least in a
dry specimen) ; the dorsal arm-plates triangular, the first two barely in contact
(when the disc diameter is 5 mm.), the rest separate ; arm-spines flattened serrulate,
five in number ; five acute oral papillae each side of each mouth angle ; adoral
shields barely meeting interradially, but with a distal lobe separating the first lateral
arm plate from the oral shield ; one large, oval, tentacle-scale throughout.

Description. The single specimen is dried and is rather twisted. Being dry
the outlines of the plates show up more than in a spirit specimen. The disc diameter
is 5 mm.; there are five arms about 27 mm. long. The disc appears to be slightly
indented in the interbrachial spaces. It is covered with large scales from most,
if not all, of which arise strong glassy spines about 0-5 mm. long ; they are thick at
the base but rapidly taper, sending off smaller spines on all sides and ending in two
or three thorns. The radial shields are moderately large, naked, convex and about
1 mm. long. They are separated from one another by a wedge of scales bearing
spines ; the wedge narrows from three scales within to one without.

i 5 2 THE "ROSAURA" EXPEDITION

The first two dorsal arm-plates are just in contact with one another, but beyond
the second one the plates become more and more widely separated from one another.
They are of a simple triangular shape, about as wide as long, with a sharp angle within,
quite straight lateral edges and a gently curved distal edge. The side-plates are
large and raised into strong spine-ridges. Those of the distal segments are very long.

There are five glassy arm-spines, increasing in length from the lowermost, which
is as long as one segment, to the uppermost, which may be as long as two segments.
They are of a characteristic shape, being flattened with each of the two blade-like
edges produced into a row of teeth. The dorsal ones taper to a blunt end but the
ventral spines have the end truncated, each corner being produced into a tooth.

Fig. 5. Ophioplinthaca grenadensis sp. n. Type. Dorsal view of the base
of an arm and part of the disk.

The proximal ventral arm-plates are just separated from one another ; those at
the end of the arm are very widely separated. They are five-sided. The two straight
or slightly concave proximal sides form an angle ; the lateral sides are deeply con-
cave ; the distal edge is widely rounded. The proximal plates are slightly broader
than long ; those further out on the arm are longer than broad.

The tentacle-scale is single and very large. It is leaf-like with a rounded free
end on the proximal segments, a more pointed end on the distal segments.

The interbrachial spaces on the ventral side are covered with large, overlapping
scales without spines. The oral shields are shaped like a broad spear-head, about
as wide as long. The proximal sides are slightly convex, the distal concave ; the
distal angle is much more broadly rounded than the proximal. The adoral shields
are tri-lobed. The long inner lobe narrows to a point which may be just in contact
with its fellow. The outer lobes separate the oral shield from the first lateral arm-
plate.

THE "ROSAURA" EXPEDITION 153

There are five or six similar triangular oral papillae on each side of the jaw and
one much larger, below the teeth, at the apex.

The colour of the preserved specimen is white.

Remarks. Having the radial shields not bar-like but oval and naked, this species
cannot be included in the genus Ophiacantha in the restricted sense. It seems to
fall within the group of genera Ophiomitrella, Ophiophthalmus and Ophioplinthaca
by virtue of the armature of the mouth-plates and the more or less distinct disc
scales.

It is distinguished from Ophiomitrella Verrill (1899a : 332) by the presence of
distal lobes to the adoral shields interposed between the first lateral arm-plates and
the oral shields. Koehler (1922 : 124) has separated off the two species which show
this character from the rest of the genus Ophiomitrella by creating a new genus
Ophiomelina. However, these two species have such insignificant radial shields that

Fig. 6. Ventral view of an interbrachial angle and the bases of the two adjacent
arms. (Some of the tentacles are drawn in.)

they are obviously not congeneric with this new species. Ophiophthalmus Matsumoto
has been limited and redefined by Koehler (1922 : 121), who relegated some
of the species to Ophiomitrella and others to Ophiomelina. None of the species
left in Ophiophthalmus has a distal lobe to the adoral shields, which instead meet
widely interradially although Ophiophthalmus hylacantha (H. L. Clark 191 1 : 227)
does have thorny tapering disc-spines, arm-spines, radial shields, ventral arm-plates
and tentacle-scales very like this new species.

Ophioplinthaca Verrill (1899a : 351), which grades into Ophiomitrella, includes
forms with oral papillae in a simple series, a single apical papilla, relatively large
bare radial shields and the disc notched interradially usually with specialized marginal
scales. Except for the last character, which is indistinct in the immature specimen
of 0. grenadensis, it conforms quite well to the other species of Ophioplinthaca. How-
ever most of these, like Ophiomitrella, have the adoral shields lacking a distal lobe
between the first lateral arm-plate and the oral shield. Ophioplinthaca partita
(Lutken & Mortensen 1899 : 179), from the west coast of Mexico, included in this
genus by H. L. Clark in 1915, differs from the rest in this respect. It seems to be

154 THE "ROSAURA" EXPEDITION

congeneric with this new species which is accordingly placed in the genus
Ophioplinthaca.

Within the genus, Ophioplinthaca rudis (Koehler) (1922 : 142, pi. 96, fig. 1) has
disc-spines which are just like those of 0. grenadensis, although their occurrence is
very variable and they may be completely absent, as illustrated by Koehler. From
Ophioplinthaca incisa (Lyman 1883), also taken at station 34, this new species differs
most obviously in having relatively smaller adoral and larger oral shields, more
numerous and spaced oral papillae, single tentacle-scales on the first pores and much
finer thorns on the arm-spines.

Fig. 7. a, Two spinelets from the disk, and b, a short blunt lower arm-spine and

a long upper one.

The genus Ophioplinthaca is also characterized by the presence of broad dorsal
arm-plates which are in contact in the proximal part of the arm of the adult. How-
ever, the relatively small plates of 0. grenadensis may also be shown by young speci-
mens of the other species of comparable size. This is illustrated in Koehler 's
photographs of Ophioplinthaca globata (1922), from the Philippine area.

The saw-like arm-spines are similar to those of Ophiacantha hirsuta Lyman (1875)
as well as to the spines of Verrill's genus Ophiopristis although that genus is also
characterized by the complex oral armature with several apical papillae.

In the present state of the family Ophiocanthidae the generic positions of many
species are open to some doubt and a thorough revision is needed to clear up the
limits of the genera.

Family Ophiactidae

Ophiactis abyssicola (Sars)

Amphiura abyssicola Sars, 1861 : 18, pi. 2, figs. 7-12.
Ophiactis abyssicola, Ljungman, 1867 : 324.

St. 49. 28 25' N., 13 34' W. (between Fuerteventura Island and Africa) ;
agassiz trawl ; c. 1,300 m. Ten specimens.

A note by Mr. Colman on the original label describes the arms as scarlet and the
body slaty-blue in life. Most of the specimens were on sponges, but it is not known
whether this is a natural association or one which came about in the trawl. The
disc of the largest is 7 mm. in diameter.

THE "ROSAURA" EXPEDITION 155

Family Amphiuridae

Amphipholis gracillima (Stimpson)

(Text-fig. 8).

Ophiolepis gracillima Stimpson, 1852 : 224.
Amphipholis gracillima, Ljungman, 1867 : 314.

St. 22. 17 28' N., 88° 11' W. (Belize Harbour, British Honduras) ; two-feet
dredge ; 6 m. Six specimens.

Most of Stimpson's original specimens had lost their discs, as three of the present
specimens have done . None has complete arms but many long fragments remain,
and where they are fractured they show no sign of tapering to an end.

Fig. 8. Amphipholis gracillima (Stimpson). Ventral view of a middle arm segment.

The species does not appear to have been described as possessing tentacle-scales.
The present specimens have two, a long narrow inner scale reaching slightly further
than the segment to which it belongs, and a small outer scale.

One specimen sent to the late Dr. H. L. Clark of the Museum of Comparative
Zoology, was compared by him with a specimen of the same size from Tobago and
with one of the types of Amphipholis gracillima. He thought there was no doubt
that the " Rosaura " specimens belong to this species, but their tentacle-scales,
particularly the inner, are bigger than in the type and the Tobago specimen. They
are accordingly figured here.

The specimens retain in spirit bars of a rich reddish colour at the distal end of
each arm-segment on the underside. |

Family Ophiochitonidae
Ophionereis dolabriformis 1 sp. n. 2

(Text-figs. 9-1 1).

St. 35. 9 25' N., 59 52' W. (off the mouth of the river Orinoco) ; otter trawl ;
86 m. One specimen.

Diagnosis. A species of Ophionereis with very fine disc-scales, becoming a
little larger around the radial shields ; supplementary dorsal arm-plates very small

1 dolabriformis — axe-shaped, referring to the ventral arm-plates.

2 Owing to delay in publication of this paper a brief description of this species has already appeared
in print. (A. M. Clark, 1953, A Revision of the genus Ophionereis. Proc. Zool. Soc. Lond. 123 : 65-94.)

156 THE "ROSAURA" EXPEDITION

and limited to the distal half of the joint, even on the proximal segments ; three long,
tapering arm-spines which vary irregularly in relative length, the uppermost usually
the longest, sometimes exceeding one-and-a-half times the length of the corresponding
joint, the middle one sometimes the shortest but often all three about the same
length ; adoral shields barely meeting interradially and also narrow outwardly,
being overlapped by the oral shield which, at its widest part just reaches the
first lateral arm-plate on each side ; disc coloured with very compact dark
reticulations.

Description. The specimen has been dried and the disc has perforated centrally
so that the teeth are visible from the dorsal side through the gap. It has contracted

2mn,

Fig. 9. Ophionereis dolabriformis sp. n. Type. Dorsal view of the base of an

arm and part of the disk.

considerably in the interradial areas so that the diameter is hard to measure ; it
is about 8 mm. All the arms are broken, the longest remaining being about 40
mm. in length and still quite stout at the end of that distance, the total length being
probably in the region of 70 mm.

The disc scaling is fine but a little coarser around the radial shields, a few of the
mid-radial scales being enlarged also. On the ventral side it is fairly uniform up to
the genital slits, which lack papillae. The scaling extends out on to the dorsal side
of the arm-bases for several segments.

The dorsal arm-plates are hexagonal in shape, the widest part being about the
middle. The two latero-proximal sides are in contact with the lateral arm-plates,
while the latero-distal sides (which usually curve round into the distal side) are

THE "ROSAURA" EXPEDITION 157

bordered by the short supplementary arm-plates, which are triangular with a curved
distal edge.

The lateral arm-plates bear three long, needle-like, tapering spines of varying
relative length ; they may all be equal, or the uppermost may be longest and the
middle one shortest, or the uppermost may be the shortest and the other two about
equal. Their maximum length exceeds one-and-a-half times the length of the
corresponding joint.

Z mm

Fig. 10. Ventral view of the base of an arm and the adjacent interradial areas.

The ventral arm-plates are shaped like a battle-axe head, with the distal edge
very convex and a little wider at the distal end of the tentacle-scales than proximal
to them. (This form might also be interpreted as " bell-shaped "). After the first
few the plates are a little longer than wide.

There is one large, oval tentacle-scale completely covering each pore.

The oral shields are broadly spear-head-shaped, the widest part being towards
the distal end and the two nearly straight proximal sides meet at an angle, while
the shorter distal sides are each slightly concave across the head of the genital slit
but form a blunt distal angle. The adoral shields are not very large and are com-
pletely overlapped by the oral shields at their widest point. Internally they just
meet inside all the oral shields except the madreporite, which is a little larger and is

158 THE "ROSAURA" EXPEDITION

swollen. There are four or five oral papillae on each side, the outermost being the
largest when they are four ; on some angles there is a small apical papilla, but on
others the lowest tooth is clearly visible. Beyond the outermost papilla is an oral
tentacle-scale.

The disc is marked with very close dark reticulations, much more compact than
those of 0. reticulata. On the arms at intervals of three or four segments there are
brownish-purple bands, each extending for one-and-a-half segments. Besides

1 mm

Fig. ii. Lateral view of part of an arm.

these there are fainter markings elsewhere on the dorsal sides of the arms, including
an ill-defined longitudinal line down the centre of the arm.

Remarks. The other three West Indian species of Ophionereis are 0. reticulata
(Say), which is very common, 0. squamulosa Koehler, known from St Thomas,
Parahiba in Brazil and common at Tobago and the Tortugas according to Dr. H. L.
Clark, and finally 0. olivacea H. L. Clark, known only from two specimens, from
Porto Rico and Florida. Dr., Clark has given a key to the identification of these
(1933 : 39). Ophionereis dolabriformis differs from the first two named above by
its small supplementary arm-plates and from all three by the needle-like arm-spines
and the coloration of the disc.

Family Ophiolepidae

Amphiophiura metabula H. L. Clark
Amphiophiura metabula H. L. Clark, 1915 : 311, pi. 17, figs. 1-3.

St. 26. 17 53' N. 87 44' W. (off British Honduras) ; agassiz trawl ; c. 900 m.
Four specimens.

St. 34. 12 05' N., 6i° 49' W. (off St. George, Grenada) ; agassiz trawl , 720-800
m. Two specimens.

The largest have discs 8 mm. in diameter. They agree in every way with Clark's
photographs except that the arm-spines are only four, or on the proximal segments
five, in number, and that they are widely separated and of minute size — less
conspicuous than in Clark's picture. Where there are four, the lowermost is close
against the tentacle-scales, the uppermost close to the dorsal plate ; the other two
are so placed that the distance between all four spines is equal, or so that the distance

THE "ROSAURA" EXPEDITION 159

between the middle two is slightly greater than that between the first and second
and between the third and fourth. Clark's fig. 3 shows at least six arm-spines, close
against one another and occupying the whole distal edge of the lateral plate

There are two specimens in the British Museum collection which were labelled
Ofihiortyfiha variabilis from " Blake " station 227, 573 fathoms (1,048 metres) off
St. Vincent. They of the same size as the " Rosaura " specimens and agree with
them in every way, including the arm-spines.

Ophiomusium validum Ljungman
Ophiomusium validum Ljungman, 1871 : 618.

St. 26. 17 53' N., 87 44' W. (off British Honduras) ; agassiz trawl ; c. 900 m.

Sixtv-five specimens. , T

The diameter of the disc is up to 10 mm., which is greater than Lyman gives.
There is a far larger specimen, diameter 17 mm. among three in the British Museum
from " Blake " station 238, 127 fathoms (232 metres) in the Grenadines Lyman
(1878 pi 1 fig. 0) shows four equally spaced arm-spines. Most of the present
specimens and some of the " Challenger " ones in the British Museum collection have
only three spines, the third (the highest) a greater distance from the second than
the second is from the first. Some have four spines at the base of the arms as the
" Blake " specimens do ; the two lowermost spines are closer together than the
second and third and the third and fourth.

Minute ventral arm-plates are present for a short distance beyond the third seg-
ment • they persist a long way out in the large " Blake " specimens.

The radial shields are usually contiguous without, but they are sometimes com-
pletely separated as in the large " Blake " specimens.

Family Ophioleucidae
Ophiernus adspersus Lyman
(Text-fig. 12)
Ophiernus adspersus Lyman, 1883 : 236, pi. 3. figs. 19-21-
St. 26. 17 53' N., 87° 44' W. (off British Honduras) ; agassiz trawl ; c. 900 m.

One specimen. . ,

St. 34. 12 05' N., 61° 49' W. (off St. George, Grenada) ; agassiz trawl, 720-800 m.

Five specimens.

The six rather broken specimens from this collection and two co-types in the
British Museum from " Blake " station 185, Dominica, 333 fathoms (610 metres)
all show a character which has not been described in this species or in any other of
the genus A single closely placed row of fine glassy bristles arises from the distal

i6o

THE "ROSAURA" EXPEDITION

Fig. 12. Ophiernus adspersus Lyman. Dorsal view of part of an arm to show
the bristles on the lateral arm-plates.

edges of the side plates on the dorsal side of the arm. In the proximal part of the
arm they are slightly longer than an arm joint ; farther out on the arm they are
longer.

REFERENCES

Bell, F. J. 1881. Description of a new species of the Genus Ar chaster from St. Helena.
Ann. Mag. Nat. Hist. (5) 8 : 440-441.

1917. Brit. Antarctic (" Terra Nova ") Exped., 1910, Nat. Hist. Rep. Zool., Echinoderma,

4 : 1-10, pis. 1, 2.

Bernasconi, I. 1 94 1. Dos nuevas especies argentinas de " Luidia." Physis. B. Aires,
19 : 117-118.

1943. Los Asteroideos sudamericanos de la familia " Luidiidae." An. Mus. argent.

B. Aires, 41 : 1-20, pis. 1-5.

Caso, M. E. 1947. Descripcion de una nueva especie del genero Moir aster de Santa Rosalia,
Golfo de California. An. Inst. Biol. Univ. Mex. 18 : 225-231, figs. 1-5.

Clark, A. M. Notes on Asteroids in the British Museum (Natural History) III : Luidia. Bull.
Brit. Mus. {Nat. Hist.), 1 (12) : 379-396.

Clark, H. L. 191 i. North Pacific Ophiurans in the collection of the United States National
Museum. Bull. U.S. nat. Mus. 75 : xvi, 1-302, text-figs. 1-144.

1915. Catalogue of recent Ophiurans. Mem. Mus. comp. Zool. Harv. 30 : 165-376,

pis. 1-20.

1917. Reports on the scientific results of the expedition to the Tropical Pacific . . .

on the " Albatross," 1 899-1 900, ... . 18. Reports on the scientific results of the
expedition to the Eastern Tropical Pacific ... by the ... " Albatross," 1904-
1905. • • • • 3°- Ophiuroidea. Bull. Mus. comp. Zool. Harv. 61 : 429-453, pis. 1-5.

1933- A handbook of the Littoral Echinoderms of Porto Rico and the other West Indian

islands. Sci. Surv. P. Rico, 16 (1) : 1-147, 1 text-fig., 7 pis.

Doderlein, L. 1920. Die Asteriden der Siboga-Expedition. 2. Die Gattung Luidia und
ihre Stammesgeschichte. Siboga-Exped. 46b : 193-293, text-figs. 1-5, pis. 18-20.

Fisher, W. K. 1906. New Starfishes from the Pacific Coast of North America. Proc.
Wash. Acad. Sci. 8 : 111-139.

191 1. Asteroidea of the North Pacific and adjacent waters. 1. Bull. U.S. nat. Mus.

76 : vi, 1-419, pis. 1-122.

1918. Notes on Asteroidea. 2. Ann. Mag. Nat. Hist. (9) 11 : 103-111.

1919. Starfishes of the Philippine Seas and adjacent waters. Bull. U.S. nat. Mus.

100 (3) : xi, 1-712, pis. 1-156.

1940. Asteroidea. " Discovery " Rep. 20 : 69-306, figs, a-m, pis. 1-23.

Grieg, J. A. 1907. Echinodermen von dem norwegischen Fishereidampfer " Michael Sars "
in den Jahren 1900-1903 gesammelt. 3. Asteroidea. Bergens Mus. Aarb. 1906 (13) :
1-87, text-figs. 1-10, pis. 1, 2.

THE "ROSAURA" EXPEDITION 161

Heding, S. 1935. The Scoresby Sound Committee's 2nd East Greenland Expedition in
1932 to King Christian IX's Land. Echinoderms. Medd. Gronland. 104(13) : 1-68,
text-figs. 1-28, maps 1-3, pis. 1, 2.

Koehler, R. 1896. Resultats scientifiques de la Campagne du " Caudan " dans le Golfe
de Gascogne. Echinodermes. Ann. Univ. Lyon, 26 : 33-127, pis. 1-4.

1908. Asteries, Ophiures et Echinides de l'Expedition Antarctique Nationale Ecossaise.

Trans, roy. Soc. Edinb. 46 : 529-649, pis. 1-16.

1909. Echinodermes provenant de campagnes du yacht " Princesse- Alice," Result.

Camp. set. Monaco, 34 : 1-317, pis. 1-32.

1922. Ophiurans of the Philippine Seas and adjacent waters. Bull. U.S. nat. Mus.

100(5) • x » 1-486, pis. 1-103.

Ljungman, A. V. 1867. Ophiuroidea viventia hue usque cognita enumerat. Ofvers.
VetenskAkad. Fork. Stochh. 23 : 303-336.

1 87 1. Forteck ofver uti Vestindien af Dr. A. Goes samt under korvetten Josefinas

Exped. i Atlantiska Oceanen samlade Ophiurider. Ofvers. VetenksAkad. Fork. Siockh.
28 : 615-658.

Ludwig, H. 1897. Die Seesterne des Mittelmeeres. Fauna u. Flora Neapel, 24 : 1-496,
pis. 1-12.

1905. Reports on an exploration off the West Coasts of Mexico, Central and South

America, and off the Galapagos Islands, ... , by the ... " Albatross," 1891,
... . 35. Reports on the scientific results of the Expedition to the Tropical Pacific,
... , on the ... " Albatross," 1 899-1900, ... . 7. Asteroidea. Mem. Mus.
comp. Zool. Harv. 32 : ix, 1-292, pis. 1-36.

Lutken, C. F., & Mortensen, Th. i 899. Reports on an exploration off the West Coasts of

Mexico, Central and South America and off the Galapagos Islands, ... , by the . . .

" Albatross," 1891, ... . 25. The Ophiuridae. Mem. Mus. comp. Zool. Harv. 23 :

97-208, pis. 1-23.
Lyman, T. 1875. Zoological results of the Hassler Expedition. 2. Ophiuridae and Astro-

phytidae. Illus. Cat. Mus. comp. Zool. Harv. 8 : 1-34, text-figs. 85-88, pis. 1-5.

1878. Ophiuridae and Astrophytidae of the exploring voyage of H.M.S. " Challenger "

1. Bull. Mus. comp. Zool. Harv. 5 : 65-168, pis. 1-10.

1883. Reports on the results of dredging, ... , in the Caribbean Sea, 1878-79, and

along the Atlantic Coast of the United States, 1880, by the ... " Blake." 20.
Ophiuroidea. Bull. Mus. comp. Zool. Harv. 10 : 227-287, pis. 1-8.

Macan, T. T. 1938. Asteroidea. Sci. Rep. John Murray Exped. 4 : 323-435, text-figs.

1-12, pis. 1-6.
Madsen, F. J. 1950. The Echinoderms collected by the Atlantide-Expedition 1945-46.

1. Asteroidea. A. F. Bruun : " Atlantide " Rep. 1 : 167-222, text-figs. 1-11, pis. 14-16.
Marenzeller, E. von. 1893. Veroffentlichungen der Commission fur Enforschung des

Ostlichen Mittelmeeres. S.B. Akad. Wiss. Wien. Math.-Naturwiss. Kl. 102 : 1-5.
Mortensen, Th. 1927. Handbook of the Echinoderms of the British Isles. London, ix + 471

pp., text-figs. 1-269.

x 933- The Echinoderms of St. Helena. Papers from Dr. Th. Mortensen's Pacific

Expedition 1914-16. 66. Vidensk. Medd. naturh. Foren. Kbh. 93 : 401-473, text-figs.
1-29, pis. 20-22.

Perrier, E. 1881. Reports on the results of dredging ... in the Gulf of Mexico, 1877-78,
by the ... " Blake," and in the Caribbean Sea, 1878-79. 14. Description sommaire
des especes nouvelles d'Asteries. Bull. Mus. comp. Zool. Harv. 9 : 1-31.

1884, Les Etoiles de Mer recueilles dans la mer des Antilles et du Golfe du Mexique.

Nouv. Arch. Mus. Hist. nat. Paris, (2)6 : 127-276, pis. 1-10.

1894. Expeditions Scientifiques du " Travailleur " et du " Talisman " pendant 1880-83

Echinodermes. Paris. 1-43 1 pp., pis. 1-26.

1896. Contribution a l'etude des Stellerides de l'Atlantique Nord. (Golfe de Gascogne,

Acores, Terre-Neuve) . Result Camp. sci. Monaco, 11 : 1-57, pis. 1-4.

162 THE "ROSAURA" EXPEDITION

Sars, G. O. 1871. Nye Echinodermer fra den norske Kyst. Fork. Vidensk-Selsk. Krist.
1871 : 1-31.

1875. On some remarkable forms of animal life from the great deeps off the Norwegian

coast. 2. Researches on the structure and affinity of the genus Brisinga. Christiania. iv -f-
112 pp., pis. 1-7.

Sars, M. 1861. Over sigt af Nor ges Echinodermer. Christiania. 1-160 pp., pis. 1-16.

Say, T. 1825. On the species of the Linnaean Genus Asterias inhabiting the coast of the

United States. /. Acad. nat. Sci. Philad. (i)5 : 141-154.
Sladen, W. P. 1889. Asteroidea. Rep. Sci. Res. " Challenger " Zool. 30 : 1-935, pis.

1-119.

1891. Report on the collection of Echinodermata from the S.W. coast of Ireland dredged

in 1888 by a committee appointed by the Royal Irish Academy. Proc. R. Irish Acad.
(3)1(5) : 687-704, pis. 25-29.

Stimpson, W. 1852. Two new species of Ophiolepis (Amphiura) from the southern coasts
of the United States. Proc. Boston Soc. nat. Hist. 4 : 224.

1853. Synopsis of the Marine Invertebrata of Grand Manan. Smithson. Contr. Knowl.

6 : 16-17.

Verrill, A. E. 1895. Distribution of the Echinoderms of North-eastern America. Amer.
J. Sci. 49 : 127-14 1, 197-212.

1899. Revision of certain Genera and Species of Starfishes with descriptions of new

forms. Trans. Conn. Acad. Arts Sci. 10 : 145-234, pis. 24-30.

18990. North American Ophiuroidea. 1. Revision of certain families and genera of

West Indian Ophiurans. 2. A faunal catalogue of the known species of West Indian
Ophiurans. Trans. Conn. Acad. Arts. Sci. 10 : 301-386, pis. 42, 43.

1 91 5- Report on the Starfishes of the West Indies, Florida and Brazil, including those

obtained by the Bahama expedition from the University of Iowa in 1893. Bull. Labs.
nat. Hist. Univ. la. 7 : 1-232, pis. 1-29.

Bull. B.M. [N.H.) Zool. 2, 6.

PLATE 6.

Fig. i

jte.A

y'-v l^/f^^ff^^^ofei

fypsijg&j^'^

,r ;* ./•".- .v. *•***-' <^r^

Mf; * k *

V f

Fig. 2

Plate 6, fig. i. Luidia rosaurae sp. n. Type.
barimae sp. n. Type. Dorsal view,

Dorsal view. Fig.
Both natural size.

2. Luidia

4 . REPORT ON THE FISHES COLLECTED BY

S.Y. "ROSAURA" IN THE NORTH AND

CENTRAL ATLANTIC, 1937-38

PART 1. FAMILIES CARCHARHINIDAE, TORPEDINIDAE,

ROSAURIDAE (NOV.), SALMONIDAE, ALEPOCEPHALIDAE,

SEARSIDAE, CLUPEIDAE

By DENYS W. TUCKER, B.Sc.,

British Museum (Natural History)
[Received 25th April, 1953]

With Plates 7 and 8 and Text-figs. 1-19

SYNOPSIS

Specimens of Carcharhinus leucas (Miiller & Henle) ; Diplobatis pictus Palmer ; Rosaura
rotunda nov, ; Salvelinus alpinus (L) ; Leptoderma macrops Vaillant ; Searsia koefoedi Parr ;
Persparsia taaningi Parr ; Pellonula afzeliusi Johnels ; Harengula pensacolae G. & B. ; and
Ilisha africana (Bloch) are reported upon. The family Rosauridae is proposed for Rosaura
rotunda nov., an aberrant new oceanic Isospondyl. The paper presents new observations and
discussion of tumours in Salvelinus ; of the structure and functions of the shoulder-organ of
Searsidae ; of the occurrence of lateral line papillae in Alepocephalidae ; and of the general
taxonomic significance of photophore systems.

INTRODUCTION

As soon as the " Rosaura " fishes were received at the British Museum (Natural
History) in 1938, a preliminary rough-sorting and assessment were carried out by
Lieut. -Col. W. P. C. Tenison, D.S.O., who also made illustrations of six of the more
bizarre deep-sea species which were included in a popular account by J. S. Colman
published as an appendix to the late Lord Moyne's Atlantic Circle in the same
year. (Another version of the same material is included in Colman, 1950, The
Sea and its Mysteries.) With the approach of war the collection was set aside and
remained unstudied until 1951, when I began my apprenticeship in ichthyology
with a very laboured consideration of it. Many of the larvae and smaller specimens
were still unnamed, and for a variety of reasons it was sometimes necessary to
modify Lieut.-Col. Tenison's determinations ; but his work had materially shortened
the initial task of identification and is gratefully acknowledged.

The collection is especially rich in well-preserved bathypelagic species from the
little-worked Caribbean and equatorial Atlantic, and the present report will lay
greater emphasis upon these than upon the shore fishes.

A general list of species taken at each station will appear with the concluding
instalment of the report, together with any minor addenda and corrigenda and
general conclusions suggested by the work. Photographs used are acknowledged
as applicable ; all other figures are from drawings by the author. Except where
specially stated, standard lengths (S.L.) of fishes are given, and denote the measure-
ment from the tip of the snout to the base of the caudal fin.

ZOOL. 2, 6 IO

164 THE "ROSAURA" EXPEDITION

My best thanks are offered to those specialists and museum curators in other
institutions whose assistance is mentioned in the text. I am especially grateful
for the interest, advice and guidance so gladly given by my colleagues Dr. Ethelwynn
Trewavas and Mr. N. B. Marshall, and for the services rendered by the technical
staffs of the Fish Section and Photographic Studio of the British Museum (Natural
History). I am indebted to Mr. J. S. Colman (now Director of the Marine Biological
Station, Port Erin, I.o.M.), who made this very fine collection and has permitted
the reproduction of extracts from his field notes which appear as quotations in
the text.

Order Pleurotremi

Family CARCHARHINIDAE

Carcharhinus sp., probably C. leucas (Miiller & Henle), alternatively
C. longimanus (Poey) or C. obscurus (Lesueur).

(PI- 7)

(i) Carcharhinus (Prionodon) leucas (Miiller & Henle).

Carcharias (Prionodon) leucas Miiller & Henle, 1841, Syst. Besch. Plagiostomen : 42. Berlin.
Carcharhinus leucas Bigelow & Schroeder, 1948, Fishes of the Western North Atlantic, 1, Sharks :
337 ; Myers, 1952, Copeia : 268.

Hab. Western Atlantic, New York to Southern Brazil ; (?) Peruvian Amazon.

(2) Carcharhinus longimanus (Poey).

Squalus longimanus Poey, 1861, Memorias hist. nat. Cuba, 2 : 338.
Carcharhinus longimanus Bigelow & Schroeder, 1948, t.c. : 354.

Hab. Tropical and subtropical Atlantic ; Mediterranean.

(3) Carcharhinus obscurus (Lesueur).

Squalus obscurus Lesueur, 1818, /. Acad. nat. Sci. Philad., 1 : 223.
Carcharhinus obscurus Bigelow & Schroeder, 1948, t.c. : 382.

Hab. Western Atlantic, New York to Florida & Bermuda ; doubtfully south-
wards to Brazil.

St. 41. 18.xii.37. °° 3 8 ' S., 43° 42' W. 2-metre stramin net, 90o(-o) metres.
One specimen (a female), 211 cm. total length. Notes and photographs taken ;
heart only preserved. Reg. No. 1953.3. 6.1.

The following data are extracted from notes by J. S. Colman :

" Shark. Female. Carcharhinus sp. (ace. to Norman). Colour — grey dorsally, white ventrally.
Two rows of teeth. No spines in front of fins. Ducti endolymphaticus present. 5 gill-slits each side.
Stomach quite empty. Ovaries very immature ; no eggs developing. I could find no parasites. Four
Remoras accompanied it."

THE "ROSAURA" EXPEDITION

'65

Dimensions

Snout to end of caudal fin (total length)

mouth (middle of upper lip)

eyes .

ist gill-slit

5th „

base of pectoral fin
ist dorsal fin

cloaca

base of 2nd dorsal fin
„ anal fin .
,, caudal fin
Girth at ist gill-slit
Max. girth (at base of ist dorsal fin)
Min. girth (at base of caudal fin)
Spread of pectoral fins

211 cm.

})

112

>»

132

134

150

113

( 6-6% T.L.)

( 7-i% „ )

(18 -9%

, )

(25-i%

, )

(24-6%

,, )

(3i-2%

, )

(53-o%

t /

(62-5%

> )

(63-5%

» /

(7i-o%

, )

On the following characters the specimen may be assigned to the family Carcha-
rhinidae in the sub-order Galeoidea : A normally shaped head, not laterally
expanded ; five gill-slits, the last situated behind the origin of the pectoral fins ;
two series of teeth functional ; two dorsal fins, which are not preceded by spines ;
first dorsal much shorter at base than caudal and terminating well before the origin
of the pel vies ; an anal fin present ; caudal fin not lunate and occupying less than
half the total length.

Indirect methods of identification, taking into account the absence of a spiracle,
the short snout, and the relative sizes, shapes and insertions of the fins, allow the
elimination of Galeocerdo, Paragaleus, Prionace, Negaprion, Aprionodon, Scoliodon
and Hypoprion. There remains only Carcharhinus, with fourteen Western Atlantic
species, of which one, C. nicaraguensis, is confined to fresh water, and others have
notably longer snouts as well as differences in the fins which are unmistakable.
Of three species one " possible," C. leucas, seems by far the most likely.

C. leucas has the reputation, consonant with a capture in a plankton net, of
being a heavy, slow-swimming species. Against an almost complete correspondence
with the smaller female described by Bigelow & Schroeder, the present specimen
offers only very small differences in the more posterior insertions of the pectoral
and first and second dorsals, and these seem negligible when considered against
small differences likely from methods of measurement and probable ontogenetic
and. other variations.

C. longimanus agrees in the short snout and long, slender pectoral, and in the
relative insertions of the fins, but has broader ist dorsal and ventral fins and longer
posterior tips to the 2nd dorsal and anal.

C. obscurus belongs to the group of species having a median dorsal ridge, the
presence or absence of which cannot certainly be determined from the photograph.

166 THE "ROSAURA" EXPEDITION

It has a short snout and other characters in common with the present specimen,
but appears to differ considerably in the shorter pectoral and lower second dorsal.
The Remoras mentioned by J. S. Colman are four examples of Remora remora
(L), 102-195 mm. S.L. In view of a very interesting paper by Szidat & Nani
(1951, Rev. Inst. Ciencias nat. y Mus. Argentino, 2 : 385) presenting evidence that
Remoras feed upon the copepod ectoparasites of sharks and other large fishes,
I made a point of examining these specimens. No recognizable organisms were
found in the very small amount of food present in their mouths and stomachs ;
the result, though disappointing, is in accord with the field-note on the absence of
shark parasites.

Order Hypotremi

Family TORPEDINIDAE

Diplobatis pictus Palmer
(PL 8)

Palmer, 1950, Ann. Mag. nat. Hist. (12) 3 : 480.

St. 35. i.xii.37. 9 25' N., 59 52' W. ; otter trawl; 86(-o) metres. One
specimen (a female), 131 mm. total length. Reg. No. 1953.3.6.2.

St. 36. 2.xii.37. 7 11' N., 57 59' W. ; otter trawl, 2o(-o) metres. Two
specimens (females), 110-140 mm. total length. Reg. No. 1953.3.6.3-4.

Hab. Guiana Coast ; mouth of R. Orinoco.

This species was previously represented in the Museum's collections by the unique
holotype, Reg. No. 1950.5. 15.4.

The specimen from St. 35 is badly deformed, extensive injuries to the body and
pectorals having been followed by irregular healing and regeneration. Those
from St. 36 (PL 8) agree adequately with the holotype, save that a rounder form
replaces the slight narrowing of the body behind the eyes, the bodies are better
nourished and the hinder edges of the pelvics have broader points. The last observa-
tion led to some conjecture whether the holotype might not be an immature male,
but there is nothing else to support that view, and by analogy with other Torpe-
dinidae recognizable claspers would be differentiated at a much younger stage.

Palmer's fig. 3, it must be noted, is very diagrammatic : the edge of the pelvic
is not smooth as shown but includes three or four shallow undulations each beset
with several small points, and anterior to the free hind tip there is a marked zone
of fusion between the inner edge of the pelvic and the ventral surface of the tail.

Reproduction of Palmer's fig. 1 rather over-emphasizes the contrast in tone
between the dark spots surrounding the light markings on the back and that of
the remaining dark blotches on the body. In the smallest specimen now available
the light spots on either side the back have an almost pearly whiteness which evi-
dently becomes much clouded in ontogeny ; also the tips of the dorsal fins are
rounder in the young stage, as in the Pacific D. ommata, and become pointed with

THE "ROSAURA" EXPEDITION 167

age. The remaining differences described between D. pictus and D. ommata (the
genotype and only other known species) are sustained by the present material.

Bigelow & Schroeder (1948, /. Mar. Res. New Haven, 7 : 43) give two main
characters separating Diplobatis from other Torpedinidae : the subdivision of the
nostril into two about the middle of its length by a bridge of stiff tissue, and a
relationship of tooth bands to the thick, fleshy lips such that the teeth are entirely
concealed when the mouth is retracted and closed.

In consideration of the possibility of ontogenetic variation in the Torpedinidae
I have examined a long series of Narcine brasiliensis (Olfers) down to embryos
with yolk-sacs attached, and find that at all stages the nostrils are undivided, so
that the condition in Diplobatis may be regarded as distinctive enough. Never-
theless the character is of doubtful use to the taxonomist, since only in the smallest
of the present specimens is the bridge intact, and even then it can hardly be described
as being of " stiff tissue " throughout ; rather does there seem to be a close ap-
proximation of a concave upper and convex lower surface with a little tenuous
tissue between the two. In the remaining material it would be difficult to decide
from superficial appearances whether a bridge ever existed. In the holotype
(of D. pictus) it remains as figured onty on the right side, and Mr. Palmer informs
me that on the left side it was ruptured during study subsequent to his paper.

Order Isospondyli

Family ROSAURIDAE nov.

Rosaura rotunda gen. et sp. nov.

(Text-figs. 1-8)

St. 42. 21.xii.37. 5 C 51' S., 34 38' W. ; 2-metre stramin net, 1200 (-0) metres.
One specimen (holotype), 8-4 mm. S.L. Reg. No. 1953. 3.6. n.
Hab. Deep Atlantic, N.E. from Brazil.

Material and methods

This solitary specimen, though but a post-larva, is sufficiently advanced in
ossification and general development to be recognized as substantially different
from all recent and fossil Isospondyli yet known.

Its transparency aroused hopes of attaining a complete knowledge of the anatcmy
of the untreated fish, but these could not be realized fully without resort to artificial
aids. I therefore made accurate notes of all structures visible in the material and
then proceeded to clear and stain, controlling all stages under the microscope and
employing, instead of the usual more elaborate methods (Hollister, 1934, Zoologica,
12 : 89-101 ; Raitt, 1935, /. Cons. Explor. Mer, 10 : 75-80) a simplification used
in the British Museum (Natural History) for many years :

168

THE "ROSAURA" EXPEDITION

-c

4fl

'43

■• — >

□

o .
Q* be
>> C

'o 5
ffi In

.: c •

C ^^
c a
u'C .£<
en rJ -*-»
^j <L> en

« -2 £

<u 2
to »-, ^5

o »>

^3 fe ^
" en Q

°s .-a

THE "ROSAURA" EXPEDITION 169

(1) Preserved material in 70% spirit transferred to distilled water — 15 minutes.

(2) Partly cleared in 0-5% KOH solution — 30 min.

(3) Stained in an old deep purple solution of alizarin in distilled water — 16£ hours.

(4) 0-5% KOH solution — 70 minutes. At this stage the specimen was still understained, so

staining was repeated with a freshly-prepared solution.

(5) Solution of about J c.c. of alizarin powder in 200 c.c. distilled water — 2 hours.

(6) Solution of 2 volumes of 0-5 KOH to 1 volume of pure glycerine. During the next 24 hours

glycerine was added a little at a time until the specimen could safely be transferred to
pure glycerine containing a thymol crystal as preservative.

The untreated specimen had sustained compound fractures of the cleithra and
wanted parts of two teeth. Processing was accomplished without further damage,
but there were several breakages of caudal and branchiostegal rays during subse-

Text-fig. 2. Rosaura rotunda. Anterior view of head from below.

quent study and manipulation, which do not seem very serious in proportion to the
additional information gained.

The text-figures were drawn freehand with the assistance of a 5 X 5 eyepiece
graticule, calibrated with a stage micrometer.

170

THE "ROSAURA" EXPEDITION

General appearance (Text-figs. 1 and 2)

Radial formula D. 16; A. 12 4-; C.23 + ; P.I2-J-; V.5(L), 2(R).

Branchiostegal rays 10. Gills 4. Myomeres of body 37.

At first glance reminiscent of the larval Ceratioids, a minute fish with an overall
length of 10-3 mm., transparent or translucent, scaleless and colourless save for a
few isolated melanophores atop the head (Text-fig. 3) and on the caudal peduncle.

Text-fig. 3.

Rosaura rotunda. Disposition of melanophores on dorsal surface of head,
shown in relation to the brain and frontal bones.

The greatest height is in the broad otic region. From this projects obliquely
forwards and downwards the narrow ethmoid cartilage. The suspensorium is
vertical, forming with the cranial axis and the jaws an approximately equilateral
triangle. Behind this the vertebral column with its myomeres descends obliquely
to the posterior end of the almost spherical body-cavity and is continued as a short
caudal region.

The body-cavity is bounded in front by an almost vertically placed pectoral
girdle, lightly curved forwards below, with the pectoral fins inserted about half-way
in the height.

The origin of the dorsal fin is substantially in advance of that of the much larger
anal. An adipose fin is present. The caudal fin is large and fan-like. A prominent
anal papilla is flanked by a pair of minute abdominal pelvic fins.

The gill-openings are wide and bordered mainly by the branchiostegal membranes,
which are continuous below (free from the isthmus), and supported by the much-
reduced opercular and subopercular and by ten conspicuous, long, slender branchio-
stegal rays. Pre- and inter-opercular bones are absent. The large mouth is
bordered by toothed premaxillaries and dentaries with prominent caniniform
teeth. The maxillaries are toothless and probably enter the gape only to a very
limited extent. The eye is small for an oceanic fish and is widely separated from
the jaws. No division can be seen in either of the single pair of minute nostrils.

THE "ROSAURA" EXPEDITION

171

Text-fig. 4. Rosaura rotunda. Osteology of head and pectoral girdle in a cleared and
stained specimen, with optical dissections of the maxillary, palato-pterygoid and
hyomandibular-symplectic-quadrate. Cartilage shown stippled. For key to abbre-
viations used see p. 181.

172 THE "ROSAURA" EXPEDITION

Cranium (Text-figs. 4 and 7)

A great part of the cranium is still unossified, and its description limited by the
difficulty of interpreting a single small and delicate specimen which may not be
dissected.

The antorbital portion of the cranium is a substantial part of the whole structure.
The largest element is the ethmoid cartilage, which is arched above, flat below, a
long cartilage extending half the length of the head and of uncertain lateral extent.
Anteriorly it is overlain by the rostral processes of the premaxillaries ; mesially
it is exposed between the widely separated anterior limbs of the frontals, and
posteriorly it sends thin supraorbital blades beneath the lateral processes of the
frontals towards the auditory capsule. The vomer and parasphenoid are applied
to its ventral surface.

The orbit is placed high on the head, laterally directed and remote from the
jaws.

The postorbital portion of the cranium is an unossified chondrocranium, globose
and including considerable auditory capsules. In dorsal view it is possible to
see, through the frontals, edges of cartilage which run parallel and close to the
optic lobes of the brain ; whether these are in fact the anterior margins of fonta-
nelles cannot be ascertained since posteriorly they become lost in the complications
of the auditor}' region, but it seems very likely that they are.

The only bones present on the dorsal surface of the head are the frontals, which
are widely separated in the middle line. They have slender anterior extensions
reaching to the tips of the premaxillary rostral processes, short posterior extensions
part over the auditory capsules, and broader lateral processes contributing to the
roof of the orbit. Separate parietals are not present.

The vomer (seen only in lateral view) is a narrow sliver of bone extending to
beneath the anterior part of the orbit, where it overlies the front end of the para-
sphenoid. The latter cannot be seen distinctly and is therefore not included in
any of the figures.

There is no mesethmoid, lateral ethmoid, or prefrontal ossified, nor are any
bones of the otic or occipital series.

Jaws (Text-figs. 4, 5 and 7)

The toothed premaxillaries form the greater part of the upper boundary of the
mouth. Each is V-shaped, the inner and longer arm of the V being a pointed
rostral process which overlaps the anterior tip of the frontal of the same side, the
outer and shorter arm being the dentigerous margin of the jaw. The premaxillaries
are separated by a narrow median fissure w T hich broadens anteriorly and posteriorly
into wide notches.

The toothless maxillaries are broadly elliptical, much like those of the Sterno-
ptychidae, and can enter but little into the gape. Each maxillary narrows anteriorly
to form a twisted process inserted into the ethmoid beneath the posterior angle of
the premaxillary. The squarish posterior end of the maxillary has a broad shallow
notch.

THE "ROSAURA" EXPEDITION

173

Meckel's cartilage persists throughout the length of the lower jaw, though almost
entirely overlain by membrane bones. The large toothed dentary is an elongated
oval with a broad notch in the wide posterior end receiving, and in part overlying,
the much smaller angular. There is a marked tuberosity at the mandibular
symphysis. The body of the angular is roughly an equilateral triangle with the
articulating facet of the quadrate above its hinder corner, and with a short narrow
limb posterior to this partly overlying the end of Meckel's cartilage. The hinder-
most tip of Meckel's cartilage is ossified as a truncated conical retro-articular.
There is no other articular ossification.

Text-fig. 5. Rosaura rotunda. Dentition, viewed from left and slightly below,
broken line shows the position of the maxillary.

The

Dentition (Text-figs. 5 and 6)

The dentition is highly raptorial. The teeth on the premaxillary are in two
series : an external row of seven minute teeth downwardly and outwardly directed
from the edge of the premaxillary, the most anterior of the series being weakly
caniniform ; and on the inner face of the premaxillary a series of four stronger,
inwardly-directed and recurved teeth. The maxillary is toothless.

In the lower jaw the dentition is dominated by a pair of very large and strongly
recurved caniniform teeth, set inside and a little behind the mandibular symphysis.
Each dentary further carries five moderately recurved caniniform teeth of oddly-
assorted sizes, and posterior to these a graded series of six to eight laterally com-

174

THE "ROSAURA" EXPEDITION

pressed, shearing teeth. So far as may be determined without manipulating the
jaws there are no vomerine teeth. (The palatines are not ossified.)

Accessory structures present in the mouth may be related to the need for re-
moving prey impaled on the enlarged mandibular teeth. The tip of the basihyal
(Text-fig. 6) bears a similar pair of enlarged and recurved teeth (one with a replace-
ment tooth behind it), evidently capable of some degree of motion between and
behind the great teeth of the mandible. Above there is a median fleshy down-
growth depending from the maxillary valve between these two pairs of teeth which
may bear some functional relation to the lower dentition.

Palatine arch (Text -fig. 4 and detail)

The palatine end of the palato-pterygoid cartilage is unossified, but is not believed
to be synchondrous with the lower end of the ethmoid cartilage. A broad ~] -shaped
endopterygoid, rounded at its anterior end and tapering to a point at the posterior,
sheaths it on its inner side. The metapterygoid is a large rhomboidal bone covering
much of the cheek between the orbit and the upper limb of the endopterygoid ;
its posterior end bears a suborbital prominence and is inserted mesiad to the hyo-
mandibular-symplectic cartilage A zone of cartilage lies exposed to the surface
at the periphery of the metapterygoid.

Hyoid arch (Text-fig. 4, detail and 6)

The hyomandibular is ossified only in its middle portion, the head and lower
end being cartilaginous. The head comprises two dorsal protuberances (unossified)
and one ventral (ossified), the latter bearing a small condyle for the articulation
of the minute opercular.

The symplectic, likewise ossified in the middle portion only, is rod-shaped. Its
upper end is synchondrous with the hyomandibular. Part of the ossified portion,
and the whole of the lower cartilaginous end, are overlain by the quadrate.

The quadrate is a short triangular ossification, the posterior edge of which is
extended upwards as a long tapering rod. The articular condyle remains carti-
laginous.

The interhyal is a short cylindrical cartilage attached at its upper end to the
cartilage between hyomandibular and symplectic and at its lower end to the lower
limb of the hyoid arch. It is not possible to distinguish with certainty between
epi-, cerato- and hypo-hyal elements, and Text-fig. 6 shows as much as can be
discerned under the limitations imposed by viewing through the suspensorium
and unopened jaws. The basihyal element is more readily distinguished and
bears a pair of enlarged teeth.

The branchiostegal rays are long, slender, tapered towards their extremities,
and elegantly curved, giving a globose contour to the anterior body of the fish.
The last four describe a somewhat S-shaped curve in traversing a vertical fold of
the gill-cover ; this curve is strongest in the uppermost (posterior) rays. The

THE "ROSAURA" EXPEDITION

175

second to sixth rays approach the hyoid arch in a smooth simple curve. The first
(lowest) branchiostegal ray is twice as broad as the rest, slightly twisted, and passes
well between the rami of the mandible. The curvature and apparent grouping of
the branchiostegal rays suggest, by analogy with other fishes, that there may be a
ceratohyal bearing six rays and an epihyal with four.

irtifu

p.mx / .

— a.

#SK

basil.

Text-fig. 6. Rosaura rotunda. Hyoid arch. The ten branchiostegal rays are accu-
rately represented ; the basihyal and interhyal are also correct, but the remaining
cartilages (epi-, cerato- and hypo-hyal) cannot be shown properly owing to difficulty
of observation. The positions of premaxillary and dentary and of some of their
teeth are shown by broken lines.

Opercular apparatus (Text-fig. 4)

The gill openings extend very nearly the full depth of the fish and are bordered
by voluminous branchiostegal membranes which are continuous below (free from
the isthmus). The branchiostegal membranes are supported by the opercular and
sub-opercular bones and the ten branchiostegal rays ; the pointed tips of the sub-
opercular and last nine branchiostegal rays project beyond the hinder edge of the
branchiostegal membrane like the ribs of an umbrella ; the tips of the opercular
and first branchiostegal ray do not reach to the margin of the membrane.

The opercular is a rudimentary, ventro-laterally directed, twisted, leaf-shaped
bone articulating with a small condyle on the head of the hyomandibular. The
sub-opercular is a narrow, pointed bone about twice as long as the preceding, con-
verging with the tenth branchiostegal ray.

Gills and gill-apparatus

Viewed as the}' have been through the suspensorium, the gill-arches have yielded
little information. There appear to be four only, bearing double rows of long and

176 THE "ROSAURA" EXPEDITION

not very numerous gill-filaments. The lower limbs of the arches bear minute
conical denticles or rakers, about a dozen to each limb.

Pectoral girdle and fin (Text-figs. 4 and 7)

The post-temporal (Text-fig. 7) is a flat, spatulate bone which mesially comes
near to meeting its fellow over the otic region of the cranium, and passes almost
transversely and horizontally before the dorsal somites to overly the upper tip of
the supracleithrum.

The supracleithrum, so far as may be ascertained in its damaged state, is a
flattened, oval bone, substantially smaller than the post-temporal ; its broad
postero-ventral end overlies the cleithrum a little below the latter 's upper tip.

The cleithrum is very large and boomerang-shaped, and with its partner comes
near to forming a complete bony ring round the body. The upper limb is nearly
vertical ; the lower is rather slenderer and curves forward and downward. The
pectoral fin is inserted behind the posterior angle of the cleithrum at about the
middle of the body-height.

The coracoid cartilage is perforated by a minute coracoid foramen and bears a
prominent, posteriorly directed ventral spine, but has no trace of ossification nor
of differentiation. Its anterior edge is inserted inside the hind edge of the cleithrum,
and there is a small an tero- ventral process which follows the edge of the cleithrum
downward.

The pectoral fin of each side is much crumpled and the structure difficult to
make out. Each appears to be a vertically-inserted reniform cartilaginous lobe,
having no actinosts ossified as yet. Twelve adult rays have stained on the upper
edge of the fin, and doubtless the number would have been increased later by the
gradual replacement of the very numerous unossified larval rays.

Pelvic girdle and fin (Text-figs. 1 and 8)

There is a very slender and apparently unpaired pelvic cartilage lying transversely
in the ventral body-wall before the anal papilla. The tapered ends of the pelvic
cartilage turn downwards and backwards into the fin-bases ; these are backwardly
directed papilliform structures from which the short close-set rays project upwards
and backwards. The fins are obviously in the early stages of development ; that
of the right side is placed higher and more anteriorly, and has but two rays com-
pared with five on the opposite side.

Axial skeleton (Text-fig. 1)

The vertebral column is as yet entirety unossified. The surface of the stout
notochord has the usual reticulate appearance ; its posterior end is curved upwards
and backwards and the slender tip projects between the bases of the 3rd and 4th
upper caudal rays. There is a slight constriction in the notochord at the point

THE "ROSAURA" EXPEDITION 177

where it breaks the edge of the fin-lobe. The first two pairs of basidorsal elements
at least are laid down in cartilage (Text-fig. 7).

The dorsal fin is supported by 16 cartilaginous basalia, each of which bears one
ray. The fin is smoothly rounded ; the middle rays are the longest, but do not
quite reach to the adipose fin when depressed. There is a wisp or two of larval
ray at the hinder end, but the full adult complement seems to be present. The
fin base is slightly elevated.

The adipose fin is rather large, slightly hooked, and with a fimbriated edge.

The anal fin, though probably incompletely developed, is still very much larger
than the dorsal. Anteriorly there is a fringed adipose blade in which further rays
might have been differentiated ; there are 12 cartilaginous basalia thus far, each
with a ray. The fin is rounded, the middle rays are the longest, and the depressed
fin reaches to the anterior rays of the caudal. The first anal ray corresponds to the
same myocomma as the eleventh dorsal ray, while the anal fin base extends beyond
the posterior end of the adipose. Discussion of verticals through fin origins would
be unprofitable, as may be observed from Text-fig. 1.

The caudal fin is supported by the tip of the notochord, and by two epurals and
three hypurals, all cartilaginous. Between the distal extremities of the hypurals
are minute triangular supplementary cartilages. There are three epaxial and
twenty hypaxial rays, all elegantly curved and tapered at their extremities ; the
broadest part of each ray lies a short distance from its insertion, as in a quill. The
dorsal rays are a little more slender and closely set than the ventral ; the rays are
evenly gradated and form a considerable caudal fan. Anterior to the caudal rays
on both dorsal and ventral edges is a fringed adipose blade in which it is evident
that a substantial number of procurrent caudal rays would have been differentiated.

Branching is incipient at the ends of most of the fin-rays.

Visceral anatomy (Text-figs. 1 and 8)

The body-cavity is spacious and near-spherical.

The funnel-shaped pharynx narrows to a short oesophagus. The stomach is
small with a median constriction. A single pyloric caecum bears subtle transverse
surface undulations suggesting subdivision into three or four at a later stage. The
intestine leaves the stomach to descend round the left side of the body, narrows as
it runs transversely upon the lower body wall, ascends some distance up the right
side, and then turns and runs obliquely downwards and backwards from right to
centre to open on the anal papilla. There is some differentiation of texture and
thickness in the regions of the gut, but no certain indication of a spiral valve nor
of any rectal glands.

The pronephros was remarkably distinct even in the untreated material, a pear-
shaped structure upon whose surface tubules are apparent, attached near the
anterior somites and the upper end of the cleithrum.

The liver comprises two major lobes, narrowing as they run dorsally to left and
right. A problematic strip of similar-seeming tissue may be pancreas ; one cannot
say.

i 7 8

THE "ROSAURA" EXPEDITION

pmx.

Text-fig. 7. Rosaura rotunda. Dorsal anatomy of the head : the lateral scale slightly
exaggerated due to drawing with a twin-objective microscope. Cartilage coarsely
stippled. The eye-muscles and auditory organ of the left side are~omitted to'avoid
over-complication. For key to abbreviations used see p. 181.

THE "ROSAURA" EXPEDITION i 79

Of the heart may be seen a large spherical auricle, a smaller thick-walled ventricle,
and a structure which must almost certainly be the sinu-auricular valve, though
the sinus itself is too transparent to be distinguished. A large vessel runs in the
mesentery to the middle of the intestine.

There is no trace of gonad development.

Musculature, etc.

The total number of myomeres developed is 37. Of these 16 are in front of the
dorsal fin ; the dorsal fin-base occupies 6 \ ; there are 2 J between dorsal and
adipose fins ; the adipose fin-base occupies 3, and 9 remain to the end of the body.
There are 22 myomeres before the anal fin, and the anal fin-base occupies 9 \. The
dorsal muscle somites are shallowest and forwardly extended where they are ap-
proximated to the posterior end of the cranium, and become progressively deeper
towards the origin of the tail. The ventral body musculature forms a thin con-
tinuous balloon-like surface, much crinkled in preservation, and with only here
and there a line to indicate a myocomma.

The m. adductor mandibulae has a superficial component inserted on the inside
of the angular in advance of the jaw-articulation ; another component inserted
nearby on the maxillary, and a deep component which turns sharply forwards and
inwards to an adjacent insertion on the inside of the dentary. The long tendon of
the m. adductor mandibulae runs along the anterior edge of the hyomandibular-
symplectic to an insertion on the head of the hyomandibular or somewhere on the
skull close by.

A very strong ligament runs from the hindermost tip of the lower jaw along the
hinder edge of the suspensorium to the region of the opercular. As in Chauliodus
(Tchernavin, 1953, The Feeding Mechanisms of a Deep-sea Fish (B.M. [N.H.]),
however, the opercular is small and the m. levator operculi, if present, consequently
likely to be weak and relatively useless in opposition to the m. adductor mandibulae.
I cannot see any ligamentous connection between the hind end of the mandible
and the posterior end of the epihyal, and therefore assume that the main work of
opening the jaws is accomplished by the m. sterno-hyoideus, which is well developed
in common with the other muscles in the floor of the mouth.

There is a patch of fibrous tissue binding the lower posterior end of the maxillary
to the dentary.

The latero-ventral boundary of the buccal cavity is formed by a thin sheet of
tissue, probably smooth muscle, whose lateral edge extends as an arc from the
level of the middle of the symplectic to the tip of the basihyal.

The abductor pinnae pectoralis is a thick muscle between the coracoid cartilage
and the anterior face of the pectoral fin ; its strong development adds another
obstruction to the observation of the radialia.

Traces of muscles are visible upon the hypurals of the caudal fan, together with
a loose aggregation of myoblasts behind the ultimate somite.

The eye muscles are long and slender (Text-fig. 7). The m.m. recti and m. obliquus

ZOOL. 2, 6. xx

i8o

THE "ROSAURA" EXPEDITION

inferior require no special comment, but the m. obliquus superior is unusual in that
it follows a very independent course to an insertion well forward on the ethmoid.

ven.

? s.a.v.

Text-fig. 8.

iv. ca

Rosaura rotunda. Abdominal viscera, viewed from a point behind and below
the vent. For key to abbreviations used see p. 181.

Nervous system and sense organs (Text-fig. 7)

Differentiation of the brain is well advanced, though there are no structures that
call for very special notice. There are olfactory lobes with olfactory tracts running
forward and outward through the ethmoid cartilage to the olfactory organs ; behind
these lie corpora striata, and between them an epiphysis or pineal organ. The
optic lobes are very large, as also is the cerebellum, and the third and fourth ventricles
are readily distinguished. The swollen origin of the spinal cord is exposed between
the anterior somites.

THE "ROSAURA" EXPEDITION

iS,

Key to abbreviations used in text-figures

ang,

angular.

oes,

oesophagus.

an.pap,

anal papilla.

o.l,

optic lobe.

aur

auricle.

olf.l,

olfactory lobe.

bas.h,

basihyal.

olf.o,

olfactory organ.

bd,

basidorsal.

op,

opercular.

brst,

branchiostegal ray.

pect,

pectoral fin.

cer.h,

ceratohyal.

pelv,

pelvic fin.

cl,

cleithrum.

pelv.cart,

pelvic cartilage.

cor,

coracoid cartilage.

pin,

pineal body.

c.s,

corpus striatum.

pmx,

premaxilla.

den,

dentary.

p.neph,

pronephros.

ecpt,

ectopterygoid.

p.tem,

post- temporal.

enpt,

endopterygoid .

pyl.c,

pyloric caecum.

eth,

ethmoid cartilage.

quadrate.

fr,

frontal.

r.a,

retro-articular.

hm,

hyomandibular.

? s.a.v,

? sinu-auricular valve

int,

intestine.

s.cl,

supra-cleithru m .

int.h,

interhyal.

sym,

symplectic.

liver.

s.op,

sub-opercular.

l.l.f,

opening of lateral line canal.

st,

stomach.

mpt,

metapterygoid.

ven,

ventricle.

mx,

maxilla.

vom,

vomer.

my,

myomere.

The eyes are laterally directed, set high on the head, and rather small for an
oceanic fish. They are loosely suspended in orbits too big for them, but there is
nothing else to evidence a post-Stylophthalmus condition. The line of suture of
the choroidal fissure is quite distinct (Text-fig. 4) .

The olfactory organs are minute and extremely difficult to interpret ; they are
undivided, and seem to have the structure of very shallow thick- walled funnels
whose openings are occluded to elliptical chinks. There is a wisp of tissue de-
pending from the edge of one which is of dubious significance.

The auditory labyrinth is very well developed ; the three semicircular canals
with their ampullae are present, and the cristae acousticae of the ampullae clearly
visible (Text-fig. 7). Part of the otolith can be seen in the bulla in dorsal view,
but its shape cannot be distinguished through the complicated light-refracting
surfaces of the auditory capsule.

Very little of the lateralis system was visible in the untreated specimen, and that
has since disappeared in clearing. There are pits with sense organs on the frontals
(Text-fig. 7) and on the dentaries (Text-figs. 2 and 4), and one or two other organs
of the suborbital and preopercular-mandibular series were faintly seen. There
is no trace of a lateralis system on the body.

Portions of the Ilnd, Illrd, IVth, Vlth and Vllth cranial nerves have been
noticed, but the observations were typical and too fragmentary to merit comment.

182 THE "ROSAURA" EXPEDITION

The surface of the body has many minute pilose projections. It is unlikely
that these indicate anything more than muscle-fibre-bundles, but since damage to
this specimen would have been of an all-or-nothing character, and in view of the
discussion of lateral line papillae in other deep-sea fishes elsewhere in this paper,
it may be worth while to put the observation on record.

Diagnosis and relationships of the Family Rosaaridae

Isospondylous fishes in which the premaxillaries have long rostral processes ;
the maxillaries are toothless but enter into the gape ; there are no supramaxillaries ;
the opercular apparatus is much reduced, the opercular being minute and the
sub-opercular but slightly less so ; the branchiostegal rays are long, slender and
relatively numerous (10 in Rosaura) and support voluminous membranes ; the
eyes are laterally directed ; the metapterygoid is large ; an adipose fin is present ;
there is an unpaired pelvic cartilage (bone ?), and there is no post-cleithrum.

One known Genus Rosaura nov. with the characters of the Family, and with
one species R. rotunda nov. based on a single immature specimen of 8-4 mm. S.L.

This definition is adequate to sustain the new family as distinct without involving
characters likely to be modified in ontogeny. Additional characters present in
the young Rosaura rotunda which will become significant if it can be shown that
they are carried through to the adult stage are : the short, stout body ; the non-
telescopic eyes ; the undivided nostrils ; the absence of pre- and inter-opercular,
pre-frontals and parietals.

In view of certain convergent similarities it may be as well to emphasize that
the Rosauridae are readily removed from the Iniomi by their possession of pre-
maxillaries which do not exclude the maxillaries from the gape ; an unpaired
ethmoid ; an unforked post-temporal ; and a reduced opercular apparatus.

Within the Isospondyli the affinities of Rosaura are difficult to trace, and it
may indeed, like Macristium, represent one of those anomalous offshoots which
are not likely to be related to any of the main branches. Perhaps its most likely
relationship is to the Stomiatoids, which group have a moderate tendency towards
the development of rostral processes, though less pronounced ; a raptorial dentition
(though usually with teeth on the maxillary, and supramaxillaries present) ; a
tendency towards loss of the parietals and post-cleithrum, and a reduced opercular
apparatus ; and a high number of branchiostegal rays (19 in Chauliodus). The
general body-form of Rosaura is not unlike that of the young stages of the Sterno-
ptychidae. But none of these resemblances is very fundamental and the differences
are formidable.

During a recent hurried visit Dr. Anton Fr. Bruun glanced at the specimen and
drawings, and expressed the belief that he had seen this little fish in some numbers
in the " Dana " Collections, and thought that a comparison with Scopelengys
tristis Alcock might be worth making. I followed up his suggestion, but from a
consultation of the literature and an examination of a specimen taken by the " John
Murray" Expedition (Reg. No. 1939.5.24.463) find that the differences are very
considerable.

THE "ROSAURA" EXPEDITION 183

Family SALMONIDAE
Salvelinus alpinus (Linnaeus)
Salmo alpinus Linnaeus, 1758, Sy sterna Naturae : 309.
Salvelinus alpinus BEPr, 1932, PBIEBI IIPECHblX BOJX CCCP H COnPEJJEJIBHBIX

CTPAH : 170 (Berg, Les poissons des eaux douces de I'U.R.S.S., Leningrad. Text in Russian,
full synonymy) ; Oliva, 1951, Copeia : 91, recent references).

St. 7. g.ix.37. 6o° 16' N., 44 41' W. ; trammel net at mouth of stream
entering Tasermuit Fjord, S. Greenland. (Pathological material.)

? Do. Three specimens, 222-330 mm. S.L. (No label.) Reg. No's. 1953.
3.6.8-10.

Hab. Arctic Boreal. (Anadromous and lacustrine races.)

If not collected on the same occasion as the pathological material, the entire
fishes must certainly be from one of the S. Greenland stations. Concerning a tumour-
bearing lower jaw J. S. Colman notes :

" When these tumours occur on the Char the fish are poor and thin and the
Greenlanders will not eat them. The authorities at Julianehaab, S.W. Greenland,
are very anxious to find out their nature and cause."

This specimen has been submitted to Prof. Alexander Haddow, Director of the
Chester Beatty Research Institute, Royal Cancer Hospital, who forwards the
following report by Dr. E. S. Horning :
" Microscopical Examination : Subcutaneous Fibroma.

" Growth consists of fibroblasts and bundles of interlacing fibrous tissue,
together with what appear to be, with this fixation," (Bouin) " elastic fibres.
There is also a rather delicate stroma of connective tissue, in which run nutrient
vessels lined by an irregular epithelium. This tumour has the appearance of
a hard fibroma, as some areas show a tendency to undergo hyaline degeneration.
It is definitely not a myxomatous type of growth, as there is no evidence of
fatty degeneration."
Another case of tumours in Salvelinus has been described by Hoshina (Jap. J.
Ichthyol. 2 (1952) : 81-88.

Family ALEPOCEPHALIDAE
(For a recent generic revision and summary of literature, see
Parr, 1951, Amer. Mus. Novitates, No. 1531 : 1-21.)
Leptoderma macrops Vaillant
(Text-figs. 9-13)
Leptoderma macrops Vaillant, 1888, Poissons Exped. Sci. " Travailleur " et " Talisman " : 166,
pi. 13, fig. 2 ; Goode & Bean, 1895, Oceanic Ichthyology : 49 (Washington) ; Koehler, 1896,
Res. Camp. Sci. " Caudan " : 523 ; Roule & Angel, 1933, Res. Camp. Sci. Monaco, 86 : 8 ;
Fowler, 1936, Bull. Amer. Mus. nat. Hist. 70 : 193 ; Bertin, 1 1940, Bull. Mus. Hist. nat. Paris
(2), 12 : 275 ; Rey, 1947, Ictiologia Iberica, 2 : 92 (Madrid) ; (?) Alcock, 1892, Ann. Mag. nat.
Hist. (6) 10 : 361.
(?) Leptoderma affinis Alcock, 1899, Cat. Deep-sea Fishes " Investigator " : 182. Calcutta ;
id., 1900, Illust. Zool. " Investigator," Fishes : pi. 32, fig. 3 ; id., 1902, A Naturalist in Indian
Seas : 237, fig. 35 (London).
(?) Leptoderma retropinna Fowler, 1943, Bull. U.S. Nat. Mus. 100 Vol. 13, Pt. 2 : 55, fig. 5.
Leptoderma Springeri Mead & Bohlke, 1953, Texas Journ. Sci. 5, No. 2 : 265 (received too late
for inclusion).

184 THE "ROSAURA" EXPEDITION

St. 49. i.ii.1938. 28° 25' N., 13 34' W. ; Agassiz trawl, c. i,30o(-o) metres.
Three specimens, 85-150 mm. S.L. Reg. No's. 1953 .3.6. 5-7.

Hab. Deep Atlantic off N.W. Africa ; Gulf of Mexico ; Gulf of Gascony ;
? Indian Ocean, Philippines. 650-2,300 metres.

The British Museum (Natural History) formerly possessed only a paratype (from
the same station as the holotype, fide Prof. L. Bertin 1 in litt.) of this species, Reg.
No. 1890.6. 16.44, still bearing the usual Paris tie-on label punched 85*223.
In all four specimens now available for study the delicate black skin has peeled
forwards, leaving most of the body naked ; as with Vaillant's other fifty-eight,
therefore, " aucun des exemplaires n'est dans un etat tout a fait satisfaisant."
Koehler (1896) likewise complains : " ... le tronc est completement pele."
Roule & Angel (1933) fared no better : " Un exemplaire, en etat defectueux

The accounts of Goode & Bean (1895), Fowler (1936) and Rey (1947) appear to
be dependent upon Vaillant's original description (1888) ; there is no evidence that
these authors handled material themselves and all perpetuate Vaillant's errors.
Koehler (1896) describes one specimen from the Gulf of Gascony. Roule & Angel
(1933) merely add a locality record. Bertin (1940) lists Vaillant's material remaining
in the Paris Museum as part of a general catalogue of types. Neglecting Alcock's
very dubious specimen (1892, 1899, 1900, 1902) and Fowler's (1943) it appears that
this species has been described only three times, and a new account may be of value.

GENERAL DESCRIPTION (BASED MAINLY ON THE LARGEST
"ROSAURA" SPECIMEN)

Radial formula D. 66; A. 86; P. 8; V.5-6; C.16.

Branchiostegal rays 6.

Gills 4 + pseudobranch.

Body naked, elongate, tapered, laterally compressed, especially in the long caudal
region, where it becomes almost filiform (Text-fig. 9).

Head small, its length equal to the trunk and 5-55 in the length from tip of snout
to caudal. Snout blunt, short, 1-27 times the interocular width and 1-40 in the
diameter of the eye ; it bears a short, blunt, laterally directed spinule on each

1 Further consideration of Bertin's " Catalogue " reveals an unfortunate misapplication of the term
" paratopotype," which affects not only the types of Leptoderma macrops but a great many others as
well. Bertin states (Bull. Mus. Hist. nat. Paris, (2) 11 : 64, 1939) :

" Les autres sont des paratypes s'ils proviennent de la meme localite que l'holotype ou des para-
topotypes s'ils ont ete pris dans d'autres regions."

The term " paratype " need not necessarily imply any such geographic restriction, while " para-
topotype " has hitherto denoted the exact opposite to Bertin's definition — c.f Alexander, 1916 (Proc.
Acad. nat. Sci. Philadelphia, 68 : 496), who introduces the term earlier in the same paper and on the
page cited defines it by implication ; also Frizell, 1933 (Amer. Midland Naturalist, 14 : 659) :

"Paratopotype " (= paratype + topotype) — a paratype from the same locality as the holo-
type."

Bertin's two categories are already covered by the terms "paratype omotype " and "paratype allo-
type " proposed by Silvestri (in Holland, 1929, Trans. 4th Inter. Congr. Entom. Ithaca : 693) ; still-born
expressions and unlamented by those who consider systematic zoology overburdened with redundant
jargon. It is a pity that " paratopotype " is as well known as it is, for few users of Bertin's invaluable

Catalogue " will feel any need to refer back to his definitions and many will thus be misled.

THE "ROSAURA" EXPEDITION

185

4->

f**tij*

tiffijj

-l->

*^r9r

*$JjjSJ

\£p

-1— 1

iy* «

■j*#'il

Cfl

• K*<S

JRufi

t/D

JMjK'

^
3

M-l

cc3

-1-H

Cfl

«o

1 86

THE "ROSAURA" EXPEDITION

side arising from the anterior end of the maxillary. Interocular space very slightly
convex, almost flat, 179 in the diameter of the eye. Eyes enormous, projecting
considerably on either side of the head, 278 in the head length. Nostrils large,
their mid-point (in the shrunken condition of the specimen) approximately midway
between the eye and the tip of the snout, the anterior smaller than and slightly
dorsal to the posterior.

Mouth small, terminal to inferior. Upper jaw 1-20 times as long as the snout,
barely reaching to anterior rim of orbit when the mouth is closed ; lower jaw 1-36
times as long as snout, reaching to below anterior eyeball. Premaxillary slender,
toothed, elongate, almost excluding maxillary from gape. Maxillary long, toothless, 1
curved, a small blunt spinule on its clubbed anterior end and a large ^-shaped
supramaxillary at the posterior (Text-fig. 10). The space between the upper ends
of the maxillary and supramaxillary is spanned by membrane, and the whole upper
jaw well shielded in a sac of thin, tough skin forming a sound functional unit.
The upper jaw rotates about its anterior end when the mouth opens and discloses
a lightly pigmented recess on the cheek. The moving maxillary is steadied against
a broad coronoid process of the dentary.

Teeth 1 in a single row in each jaw, about 14-20 on the premaxillary, seldom

Table I. — Leptoderma macrops : Table of Measurements and
Comparative Proportions

It should be noted that Vaillant writes variously " Longueur " and
In the present case he gives the length of the caudal fin as 3 (?) mm.

Longueur totale."
the likely error in

a false assumption of standard length is therefore of the order of 0-1-0-4% underestima-
tion in the figures quoted, and may be neglected.

Total length .....
Standard length ....

Head length .....
Tip of snout to pectoral fin
,, ,, pelvic ,,

,, ,, vent

,, ,, anal fin .

,, ,, dorsal fin .

Greatest depth of head

body behind head .
thickness of body behind head
Depth at vent ....

Length of pectoral fin
Longest ray of dorsal fin .
,, ,, anal ,, .

Least depth of caudal peduncle

Snout

Eye

Interorbital width .

Interocular , ,

Upper jaw

Lower ,,

Greatest depth of head

,, breadth ,, (across eyes)

Rosaura specimen
152 mm.
150

27 mm.

12-4

9-7

12-2

9
0-7

6-9 mm.

9'7
2-4

5-4
8-3
9.4

12-4
14*6

Holotype
1 Longueur
164 mm."

Percentages of standard length

18 -o%

2i'3%

32-6%

36-6%

38-o%

5o-o%

8-2%

7-3%

5-3%

6-4%

8'i%

o-4%

25-5%

35'9%

8-8%

20 -o%

3o-7%
34-8%
45-9%
54 -o%

17-6%

ca. 37'5% (3/8)

7%
2%

Percentages of head length
27.5%
37'9°o
17-2%

1 I am unable to find the " single row of minute conical teeth " at the anterior end of the maxillary,
described by Mead & Bohlke (1953). But theirs is a larger specimen (190 : 150 mm.).

THE "ROSAURA" EXPEDITION 187

extending far to the side, and about 12-15 on the dentary. The teeth in each jaw
are simple, slightly incurved, readily depressible fanglets (Text-fig. n), and do
not exceed 0-2 mm. in length in the largest specimen available. There is a danger
of being misled into believing that the dentition is multiseriate, because rows of
white structures (believed to be taste-buds rather than photophores) stand imme-
diately behind the teeth, as well as being lavishly scattered over the lips (Text-fig.
n), and buccal cavity. The teeth may be largely concealed behind the lips ; this
condition varies in different specimens. There seem to be no palatine or vomerine
teeth.

Gill-opening wide, though only extending upward to the level of the middle of
the eye ; membranes voluminous, posteriorly separate and free from the isthmus.
The six branchiostegal rays and the opercular bones which together form the gill-
cover are all concealed under heavily-pigmented skin, so that only the more promi-
nent ridges of the bones show white at the surface and their boundaries are largely
concealed. The preopercular has a long elbowed ridge extending from a point
midway between the eye and the top of the gill-opening to the articulation of the
jaw. The interopercular accounts for another ridge arising in the angle between
pre- and sub-opercular and running some way with the former. The opercular
has about a dozen ridges, some of which terminate in small denticulations at the
hinder margin. The subopercular has one principal ridge which supports a small
lobe at the lower posterior corner. The branchiostegal rays are slender and increase
in length posteriorly ; the sixth appears (without dissection) to be a rather stouter
structure of T-shaped cross section (Hubbs' " branchioperculum ").

Gills 4 ; a pseudobranch of half-a-dozen filaments present on the inside of the
gill-cover. Gill-rakers 12 on lower limb of first arch ; lanceolate with a few minute
bristles. Gill-filaments rather short, leaf-like, shorter than rakers.

Pores of lateralis system of head few and large ; 3 in the supra-temporal series,
3 or more in the supra-orbital, 4 in the infra-orbital, 5 in the preoperculo-mandibular
(Text-fig. I2A). There may be others on the snout.

The thin skin of the body has been stripped forward and hangs in tatters about
the head ; it is possible, however, to paint a few centimetres of it back into position
with a fine brush, and to decide pretty certainly that there were no lateral line
pores on the body. Except that the skin tends to fold into a groove of uncertain
significance neither is there any sign of a lateral-line canal upon the skin, nor upon
the horizontal septum of the body. There are, however, upon the head and along
the skin of the body large numbers of " tag organs " or lateral line papillae, in
tracts corresponding to the theoretical lateralis distribution although quite possibly
independently innervated (Text-fig. 12 A, b). These papillae are compressed,
closed, and of much the same shape and variability of form as the common littoral
sponge Grantia compressa. The papillae of the body (Text-fig. 12c) and operculum
are longest, and approach 0-3 mm. ; they are heavily pigmented save at the tip
and in a very narrow marginal zone, and have a core of connective tissue which
is in some cases seen to be penetrated by a short branch of a nerve running along
the papilla-tract. It is presumed that sensory cells were located in the tips of the
papilla, but surface magnification reveals none, and since the fishes have taken a

THE "ROSAURA" EXPEDITION

I MM

0-3

Text-fig. 12. Leptoderma macrops. a. Diagram showing disposition of lateral line
pores and sensory papillae over the head and trunk, b. Three papillae of the sub-
orbital series, and a papilla and a pore of the preoperculo-mandibular series. (The
area enclosed in the square on a, more highly magnified.) c. A single sensory papilla
from the body. (That arrowed in a, more highly magnified).

great deal of punishment in the trawl the chances of their preservation are con-
sidered too slight to justify sectioning for the present.

Upon the head the papillae become shorter and squarer, until on the tip of the
snout and about the lips they become very small, lose their pigment, and become
superficially identical with the presumed taste-organs inside the mouth. (See
further discussion on p. 193.)

The vent opens on a slight prominence exactly three-eighths the standard length
from the tip of the snout.

Pectoral fins long, narrow, obliquely inserted about one-fifth the body-height

THE "ROSAURA" EXPEDITION 189

from the ventral surface. All the rays except one are deeply forked, nearly to
their bases.

Pelvics ventral, abdominal, separated by the width of one base, reaching to
the anal.

Dorsal and anal fins very long, extending almost to the caudal, and each erected
fin much taller than the body depth at the same level. The anal commences
immediately behind the vent ; the dorsal much further back at the level of the 15th
anal ray.

Caudal with a distinct and very slender peduncle, ending in a rather prominent
little fan-like expansion of hypurals. The caudal rays are much damaged, and
both the length and the count likely to be underestimated.

Colour (in spirit) deep purple-black all over the body, inside the buccal cavity
and gill-chamber (and probably inside the body-cavity) ; considerably lighter on
the snout tip and eyeballs. There is no evidence of organized photophores, though
this does not preclude general luminescence of the epidermal mucus.

The largest specimen to hand is a female, with opaque white eggs of varying
diameters (0-4-1-2 mm.) extruding from the vent. The extrusion is probably a
decompression phenomenon, although Vaillant has already commented upon the
great variation in the sizes of the eggs in this species.

COMPARISON WITH VAILLANT'S DESCRIPTION (1888)

Vaillant's description is comprehensive, detailed, and for the most part more
accurate than most work of his period. Despite the few marked discrepancies
now to be noticed there could be no reasonable doubt but that the present material
should be referred to Vaillant's L. macrops, even without the advantage of a para-
type for comparison. The new account is more an expansion of Vaillant's than a
correction, although paucity of material leaves his account of the internal anatomy
unchecked.

Vaillant gives D.48, A. 71 +, and his figure agrees, showing these fins extending
nearly to the caudal. The paratype now to hand, a specimen of 123 mm. S.L.,
gives practically these counts in median fins ending abruptly more than a centimetre
before the caudal, and beyond their terminations there remains little more than
vertebral column and a few tattered caudal rays. Having regard to the spacing
of the posterior rays, and assuming that the fins once reached nearly to the caudal,
as they do in Vaillant's figure and in the " Rosaura " material (which, it must be
noted, was collected not far from the type-locality and at a similar depth), I estimate
D.46 + ca. 20, A. 74 -f ca. 12, a precise correspondence with the "Rosaura"
specimens. It is possible that Vaillant's artist reconstructed the dorsal, anal and
caudal in their correct relations, perhaps from a better specimen, and fitted in the
rays in accordance with counts supplied by Vaillant himself from a different fish.

Later in his text Vaillant gives P. 15 (uncountable from the figure). Though at
first I agreed, counting from the projection of the fin through a hole in the loose
skin, I later found that I had merely repeated his error in counting the branches of
rays which bifurcate close to their bases. Of the ventrals Vaillant remarks only

i 9 o THE "ROSAURA" EXPEDITION

that they are " si rapprochees qu'elles semblent se confondre " ; the figure appears
to show V . 5 or 6 and I count V . 5 on the specimen.

The measurements available and shown in the comparative table (p. 186) show
close agreement, except at one point where it seems evident that Vaillant's " inter-
orbital " corresponds with my " interocular." He gives B.V + instead of B.VI ;
he slightly undervalues the length of the maxillary ; he mentions the dorsal fin
originating over the 23rd anal ray instead of the 15th — a likely variation and one
common in other elongated fishes, such as Notacanthus for example ; and he fails
to notice the sensory papillae, which, with the condition of his material and the
probable magnification used in examining his specimens, is understandable enough.
(Vaillant seems to have been impressed by the minuteness of the teeth — " assez
peu developpees, pour n'etre visibles qu'a un assez fort grossissement " — and the
papillae of the lips and snout are smaller still, while, having decided that the lateral
line was absent, Vaillant might well have felt no need to examine the skin of the
body.)

COMPARISON WITH LEPTODERMA AFFINIS ALCOCK, 1899

Alcock (1892) described as L. macrops a single fish of 8 J inches (222 mm.) S.L.,
trawled by the " Investigator " at 753 fathoms (1,357 metres) in the Bay of Bengal ;
later (1899, 1900, 1902) he published a fuller description and figured it as L. affinis.
Mr. A. E. Parr (Director of the American Museum of Natural History, at present
engaged on a monograph of the Alepocephalidae) tells me that on inquiry of the
Indian Museum, Calcutta, recently he learned that Alcock's unique holotype had
been lost. No other records of Indian Ocean material have been given and Alcock's
papers therefore provide the only evidence for consideration.

The invariably damaged condition of L. macrops in collections has been empha-
sized ; L. affinis stands in striking contrast, although the conditions of its capture
were not particularly auspicious. Alcock makes no mention of the skin peeling
from the body and is able to describe the lateral line in its entirety ; thus (1899) :

... a row of pores extends from the occiput to the caudal. In spirit
the colour is purple, the contracted opaline epidermis forming a sort of bloom."

Again, the popular account (1902) :

... it is quite black and has no special phosphorescent glands, yet its
entire skin is enveloped in a thick, opalescent epidermis, like a luminous
' bloom.' The only specimen captured glimmered like a ghost as it lay dead
at the bottom of a pail of turbid seawater."

We can be satisfied as to the perfect condition of Alcock's specimen, and wonder
indeed whether its apparent resistance to abrasion in the trawl may not indicate a
specific difference from the Atlantic form. Alcock (1899) concludes :

" This species seems to differ from Leptoderma macrops Vaillant in having
the body less elongate, the lateral line very distinct, and the rays of the dorsal
and anal fins more numerous. It agrees fairly well with the figure but not
with the description of that species, and is probably identical with it."

THE "ROSAURA" EXPEDITION 191

The revised fin-ray counts for L. macrops show a remarkable similarity to those
for L. affinis. Adding pectoral and caudal counts taken from Alcock's figure (1900)
to the incomplete formula published in 1899, we have :

L. affinis T).ca. 66) A.ca. 85 ; P. 8; V.5 ; C.17
L. macrops T>. 66; A. 86; P. 8; V.5-6; C.16

Despite this similarity there remain several serious discrepancies between the
two species.

(1) Body proportions. — From Alcock's publications it may be deduced that in
L. affinis the body-height measures 12-6% and the head-length 25-7% in a standard
length of 222 mm. (Alcock does not state whether " standard " or " total " length
is intended in this case, but his frequent mention of " length without caudal "
elsewhere in his report of 1899 sufficiently indicates his custom. The proportions
given, being taken from his figure, are unaffected by this consideration). Com-
parative figures are 67%, 17-6%, 164 mm. for Vaillant's holotype, and 7-3%,
18%, 150 mm. for the largest " Rosaura " specimen of L. macrops.

It would be hard to attribute these wide differences to the effects of ontogenetic
changes, geographic variation and the fortunes of preservation. There is in fact
evidence that the relative size of the head decreases with age in L. macrops, whereas
the identification of L. affinis with L. macrops would require a contrary tendency.

L. macrops

L. affinis

" Rosaura " Paratype " Rosaura " Holotype (Alcock)

Standard length (mm.) . . 85 . 123 . 130 . 150 . 164 . 222

Head as % S.L. . . . 20-5 . 19-5 . 18-4 . 18-0 . 17-6 . 25-7

One further explanation may be considered. The caudal peduncle in Alcock's
figure is short and deep ; despite the disarming elegance of the caudal fork is it
possible that the figure represents a truncated and regenerated tail ? If so, then
estimates of the length of the missing portion based on computations of the length
necessary to reconcile the proportions of L. affinis with those of L. macrops should
show reasonable agreement.

Estimated

L. macrops

L. affinis

truncation.

Standard length

150 mm.

222 mm.

—

Head length
Snout to anal
Snout to dorsal

18 -o%
38-o%
50-0%

57 .»
105 „
129 ,,

94 mm.
65 „
36 „

This hypothesis is clearly untenable also.

(2) Lateral line. — Alcock comments upon the apparent difference between his
species and Vaillant's in this respect ; it was a point which he would undoubtedly
have verified. I have examined the skin and bodies of four specimens of L. macrops
and find no pores save on the head ; it seems hardly likely that the stripping of the
skin should destroy them so thoroughly as to defy microscopic examination. Yet
is Alcock truly representing the case when he speaks of " a line of pores " ? His
figure suggests, rather, a line of sense organs lying flush with the skin.

i 9 2 THE "ROSAURA" EXPEDITION

(3) Sensory papillae. — Nineteenth-century workers 1 did not often use very high
magnifications when examining fishes, and Alcock may have overlooked papillae
for this reason, as did Vaillant.

(4) Supramaxillary . — I find one supramaxillary in L. macrops, as also did Vaillant
— " un sus-maxillaire distinct," " un petit osselet en serpe." Alcock (1899) on
the other hand speaks of a maxilla which " is very broad and consists of three
pieces." It is unthinkable that he should have included the entire premaxillary
in this trio, but very likely, as his figure seems to indicate, that he mistakenly
dissociated the posterior process of the premaxillary from that bone as a second
supramaxillary.

(5) Pectoral fin. — The pectoral of L. macrops is low on the body and almost
horizontally inserted. Alcock's figure shows that of L. ajfinis much higher and
practically vertically inserted.

The evidence seems at the very least to support the retention of Leptoderma
ajfinis Alcock as a distinct species. If Alcock reported correctly the presence of
two supramaxillaries, moreover, we are faced with a distinction of generic importance,
and one which Parr (1951) uses as the basis of the first dichotomy of his key to
the genera of Alepocephalidae. Conversely, with the retention of L. ajfinis in
Leptoderma, the key drafted by Parr will require modification of the stated body-
proportions, as well as of the fin-ray counts corrected by the present work.

BIONOMICS AND ADAPTATIONS OF LEPTODERMA MACROPS

" ... the ' Michael Sars ' at Station 48, between the Canaries and the Azores, brought up an
Alepocephalus in the large trawl towed at the bottom in 5,000 metres, just as these fishes have been
captured by most deep sea expeditions ; on the trawl-rope a small tow-net was fixed in such a way
that it was towed about 1,000 metres above the bottom and in this net an Alepocephalus was also
captured. Such facts warn us against hasty conclusions.

Murray & Hjort, 191 2, The Depths of the Ocean : 412. London.

Leptoderma macrops has (on published records) been taken in ten trawl hauls ; by
the " Talisman " (6), " Caudan " (i), " Princesse Alice II " (i), " Oregon " (i), and now
the " Rosaura " (i), in depths ranging from 650-2,330 metres. Of a total of 65 speci-
mens 59 were collected by the " Talisman ", 47 at St. 93 on the Banc d'Arguin. This
paucity of records is remarkable when one considers the number of bottom hauls
made by, say, the " Michael Sars," within the known range of the species and without
results. Unless the " Talisman " encountered a breeding concentration the fish
seems singularly capricious in its occurrence.

The association of predominantly bottom fauna with L. macrops is indicated in
the reports of Vaillant (1888) and of Roule & Angel (1933). At " Rosaura " St. 49
it was accompanied by 2 Bathypterois dubius Vaillant (with minute eyes and tactile
pectorals) ; 5 Gadomus longifilis (G. & B.), 1 Gadomus arcuatus (G. & B.), 5 Bathy-
gadus vaillanti Roule & Angel, 1 Halosaurus oweni Johnson, 1 H. johnsonianus
Vaillant and 29 Synaphobranchus pinnatus Gronov. In anticipation of the in-
vertebrate papers, and by courtesy of Dr. M. Burton, Mr. N. Tebble, Dr. I. Gordon,

1 Charles Darwin, for example, had a great preference for the simple microscope, and had no com-
pound microscope with him on his " Beagle " voyage (Life and Letters, ed. F. Darwin, 1 : 145. London,

THE "ROSAURA" EXPEDITION

193

Dr. W. J. Rees and Miss A. M. Clark, it is possible to report associated bottom fauna
in the groups Porifera, Polychaeta, Decapoda, Mollusca (dead pteropod shells)
and Echinodermata. Preliminary assessments indicate that no mid-water species
were taken in any group.

All the available evidence therefore points to Leptoderma macrops being a fish
living on the bottom beyond the Continental Shelf. That it is of the select group
of deep-sea fishes which perform considerable vertical migrations is unproven and
unlikely : its general appearance makes this seem much more improbable than in
the case of Alepocephalus macrolepus Koefoed (1910, Set. Rep. " Michael Sars,"
4 : 44), the fish whose exploit is recorded in the cautionary text at the head of this
section.

Little can be learned from the gill-rakers. The filaments suggest that cutaneous
respiration has permitted a reduction in their size. The small gape, puny dentition
and unathletic form point to a fish living upon detritus and /or small and feeble
prey. The jaws are not protractile, though the maxillaries are practically vertical
when the mouth is fully opened, so that in one way the following comparison is
invalid, but the great number of taste-organs in the mouth may point to the fish
living like a deep-sea carp, sucking -detritus from the bottom and sampling it for
nutritious particles. (I have not gutted the specimens, for obvious reasons, but a
radiograph of the largest shows objects in the stomach which resemble a small
lamellibranch shell and a couple of large sponge spicules.) The system of lateral
line papillae, obviously useful in warning of enemies, may be used offensively against
small prey such as polychaetes and crustaceans, a vibration in the water being
followed by a reflex snap and a leisurely assessment of the catch by the taste organs.
(See also p. i94#.) In either case Walls' dictum holds : " Go far enough along the
bottom (if you're a fish), and you're bound to bump into something good to eat."

The excellent summary of the eyes of deep-sea fishes by Walls (1942, " The
Vertebrate Eye," Bull. Cranbrook Inst. Sci. 19 : 391-405) is the starting point for
the discussion which follows. Vaillant (1888) has admitted that his published
figures do not adequately represent the lateral protrusion of the eyes of Leptoderma.
The diagrammatic reconstruction presented in Text-fig. 13 attempts to do this,
and, it is believed, with considerable accuracy. The largest " Rosaura " specimen
was used for it, the left eye being undamaged, and in an apparently normal position
in the orbit, and the shrunken cornea was gently pulled out to its proper convexity
with fine forceps. The lens is not shown ; in dorsal view it is concealed under the
pigmented portion of the eyeball, and whether this condition is normal or not it is
certainly very large, much larger than the diameter of the pupil.

These eyes are enlarged, but quite unlike those of the fishes of the twilight zone,
such as Epigonus and Aphanopus. If Walls is correct in stating that the eyes of
Bathytroctes and Platytroctes are " apparently in a half-way stage of evolution
toward an eventual forward-aimed, tubular organ " like those of Gigantura and
Winteria, then it must be admitted that Leptoderma is a great deal more advanced
towards the same type. The parallel tubular eyes already cited are, like the
upwardly directed ones of Argyropelecus and Opisthoproctus, most likely concerned
with space-perception, assessing the distance of the infrequent flashes from the

194

THE "ROSAURA" EXPEDITION

photophores of other organisms. With a light penetration limited to 400 metres
depth it is most unlikely that Leptoderma can ever aspire even to the twilight zone ;
the structure of its eyes makes it seem most unlikely that it can form a distinct
image under whatsoever lighting conditions, and yet it is clear from Text-fig. 13
that if light-sensitivity be admitted as " vision " Leptoderma has, by moderate
nasad asymmetry and considerable protuberance of the eyes, achieved a very fair
degree of stereoscopic vision, if not of binocularity. If its eyes have largely lost
their effectiveness as cameras through structural modification and conditions of
intense darkness (and it will be noted that the small ocular photophores concerned
with neuro-facilitation in fishes which rely on normal vision below the photic zone

Text-fig. 13. Leptoderma macrops. Diagrammatic reconstruction of dorsal surface of
head, showing lateral projection of the eyes.

are not developed in this genus), it is still obvious that they have a high degree of
potential usefulness as range-finder photometers, capable of locating animal and
bacterial luminescence and enabling assessment of its potentialities as food or foe.

The associated bottom fauna from St. 49 included numerous specimens of Poly-
cheles spp. and Heterocarpus sp., crustacean genera whose Indian Ocean species at
least are known to be luminous, and also the Asteroid Brisingella coronata which
belongs to a family whose luminescence is in dispute. It is also interesting to note
that in this one haul the enlarged eyes of Leptoderma macrops are paralleled in a
polychaete Robertianella and a crustacean Heterocarpus, while the reduced eyes of
Bathypterois dubius are parallelled in the Crustacea Nephropsis atlantica and Poly-
cheles spp.

OCCURRENCE AND SIGNIFICANCE OF SENSORY PAPILLAE
IN DEEP-SEA FISHES

Papilliform lateral line organs have been reported in several quite unrelated
groups of deep-sea fishes. Zugmayer (191 1, Result. Camp. Set. Monaco, 35 : 92,
pi. iv, fig. 3) records them among the Lyomeri in Gastrostomus (= Eury pharynx).
Pfuller (1914, Z. Naturw. Jena, 52 : 1) in an intensive study of the lateral line organs

THE "ROSAURA" EXPEDITION

195

of the Macruridae gives accounts and figures sections of papillae in Coelorhynchus
and Macrurus. Roule & Bertin (1929, Dana Reps. No. 4 : 12) mention the co-
existence of lateral line pores and papillae in the deep-sea eel Nemichthys, and the
same authors (op. cit. : 53) and Trewavas (1933, Proc. zool. Soc. London : 601)
describe a variant of the same condition in the nearly-related Cyema. The most
comprehensive demonstration of papilla variation is given by Regan & Trewavas
(1932, Dana Reps. No. 2 : 23), who describe and figure five distinct types in the
Ceratioid Angler-fishes, and show that the several families may be classified into
five corresponding groups. Their Dolopichthys papilla approaches most nearly to
that of Leptoderma.

Among the Alepocephalidae, however, no very precise demonstration seems to
have been given before the present. In the Alepocephalidae sensu lato (including
Parr's Searsidae), Brauer (1906, Wiss. Ergeb. Deutsch. Tiefsee-Exped. " Valdivia,"
15 : 17 et seq.) describes " papillen " in several species, with recognizable figures,
but makes no functional distinction between photophores general and special, the
" shoulder organs " of Searsidae, and true lateral line papillae. The species
mentioned by Brauer, in none of which seems a Leptoderma condition to be clearly
demonstrated, are :

Fam. Alepocephalidae

Bathytroctes longifilis (= Nemabathytroctes longifilis (Brauer)).
Aleposomus lividus (= Rouleina lividus (Brauer)).
Aleposomus nudus (= Rouleina nudus (Brauer)).

Fam. Searsidae

Bathytroctes rostratus (nee Giinther) (= Searsia koefoedi Parr).

Beebe (1933, Zoologica, N.Y. 16 : 17) describes a pattern of tubercles in his
Anomalopterus megalops, but which can hardly be reconciled with a lateralis dis-
tribution (nor, for that matter, Beebe's species with Anomalopterus Vaillant (now
Anomalopterichthys Whitley, the name being several times preoccupied). Beebe
(op. cit. : 83) also describes small stalked structures in Photostylus which he regards
as photophores, but which may well be primarily sensory papillae ; these also are
not regularly arranged.

Unfortunately many genera of Alepocephalidae are not yet represented in the
collections of the British Museum (Natural History) but a rapid search for other
forms bearing lateral line papillae has been rewarded. The holotype of Xeno-
dermichthys nodulosus Giinther (1887, Set. Rep. " Challenger " Zool. 22 : 230, pi. 58,
fig. c) is much as Giinther described it, with a normal tubular lateral line, and a
tough skin scattered with minute white scale-like structures and small, nodular
and undoubted photophores. Somewhat surprisingly Giinther failed to notice
that almost the whole of the head (including the isthmus) is unpigmented (save
for the photophore rims), intact, and sharply distinguished from the body ; a few
years later and he must surely have noted an appearance strongly recalling the all-
luminous head of Aulastomatomorpha. He also missed the lateral line papillae :
minute, club-shaped, with slender colourless attachments and black tips (Text-

ZOOL. 2, 6. 12

ig6

THE "ROSAURA" EXPEDITION

figs. 14 and 15). They are arranged in single and double linear tracts upon the
head, corresponding to the lateralis distribution, and are upon the body rather
sparsely placed immediately above the lateral line tube and upon the tube itself,
either between or above the lateral line pores. Down the upper limit of the gill-
cover, along the boundary between pigmented and unpigmented integument, the
papillae are more elongate and their stalks whiter ; they resemble photographic

tti*

• . • :.. . *"** ■ • •. • ^T •. . .• >^-"y •••■

PA
PO

5 MM.

Text-fig. 14. Xenodermichthys nodulosus Giinther. (Holotype, 180 mm. S.L.).
Portion of skin from the left flank. The lateral line, ll, crosses the centre of the
figure, po, lateral line pore ; pa, papillae ; s, rudimentary scale, three of which
may be seen supporting the outer wall of the lateral line canal from within ; ph, large
photophores, arranged in alternating segmental rows so as to mark out rhomboidal
areas on the body ; mp, very numerous white spots, possibly luminous (mucus glands
according to von Lendenfeld), arranged in close longitudinal rows.

negatives of the fungus Xylaria hypoxylon. Von Lendenfeld (1887, " Challenger "
Reps. Zool. 22, Appendix B. : 307, pi. 53, figs. 49-53) has described the histology
of the photophores and mucus glands, but again does not mention sensory papillae.
It is likely that he was given a piece of tissue from the flank (missing in the speci-
men) which would not have borne any papillae, and that he did not receive the
entire fish for examination. Rauther (1927, Br onus Tierreich, 6, Abt. 1, Book
2 : 125) in his discussion of light organs in fishes regards the first-mentioned struc-
tures as sensory organs.

Convergent evolution of lateral line papillae in several widely separated groups
of deep-sea fishes has not yet received an explanation, despite its clear adaptive

THE "ROSAURA" EXPEDITION

197

value. All the fishes concerned are forms of inconsiderable locomotive powers,
not subject to violent water movements consequent upon their own swimming,
and living at depths beyond the range of wave action and of strong currents. In
such conditions extreme exposure of lateral line sense organs upon papillae results
in no redundant irritation to the fish ; on the contrary, it must materially assist
the apprehension of the slightest local disturbance in the water, and be a vital
asset in the collection of information from an environment that can impart so

0-5 N\M.

Text-fig. 15. Xenodermichthys nodulosus. Three types of sensory papillae, a. Stout
papilla from occipital region, at the boundary between pigmented and unpigmented
integument. The papilla bears two large photophores towards its base, and has a
number of minute, intense white spots scattered over its surface, probably further
photophores. b. Clavate papilla standing upon the supra-orbital canal. The papilla
is traversed by a nerve, and has presumed minute photophores over its surface. This
is a type which punctuates rows of similar papillae of about half the height in the
lateralis system of the head. A similar size and type of papilla to this stands upon
and above the lateral line canal of the body. c. Slender-stalked papilla, type confined
to upper margin of operculum, along boundary between pigmented and unpigmented
integument. (Text-figs, a, b and c are drawn to the same scale.)

little so reluctantly to the more orthodox sense organs. It is probable that the
lateralis system in these fishes allows precise location and capture of small prey
without any visual assistance. Experimental evidence in support of this theory
may be derived from the experiments of Dykgraaf (1932, Z. vergl. Physiol. 17 : 802),
who showed that the freshwater minnow Phoxtnus can be trained to snap at small
bodies gently agitated in the vicinity of its caudal lateral line papillae ; in this
case, however, the sensory cells are situated at the bases of the tubular papillae,
instead of at the tips as in the deep-sea fishes so far investigated. For a general
introduction to this subject see Pf tiller (ante) and Wunder, 1936 (in Demoll & Maier,
Handbuch der Binnenfischerei Mitteleuropas, 2B : 49. Stuttgart).

i 9 8 THE "ROSAURA" EXPEDITION

Family SEARSIDAE.
Parr, 1951, American Museum Novitates, No. 1531 : 1-21.

ALEPOCEPHALIDAE (part)

Gunther, 1887, " Challenger " Reps. Zool. 22 : 227-8.

Brauer, 1906, Wiss. Ergeb. deutsch. Tiefsee-Exped. " Valdivia," 15 : 17-18.
Beebe, 1933, Zoologica N.Y. 16 : 36-56.
Parr, 1937, Bull. Bingham Oceanogr. Coll. 3 : 12-19.
And numerous other authors.

NOTES ON TAXONOMIC AND STRUCTURAL FEATURES

(1) Shoulder organ

The Searsidae have been separated off from the Alepocephalidae by Parr (1951),
and are characterized by the possession of a peculiar structure which Parr, in this,
his latest paper, calls the " shoulder organ," which name is the least unsuitable

>x -_ lettered line

Supraclavicular

process

Text-fig. 16. Diagram of shoulder organ (" supraclavicular process ") in a young
Searsid, showing its relation to pectoral girdle and lateral line, pto, post-temporal ;
pel, supracleithrum ; cl, cleithrum. (From Beebe, 1933.)

of those so far applied. The new family forms a quite distinct group, and its recogni-
tion seems surprisingly overdue, although one regrets its immediate proliferation
of new subfamilies, genera, and subgenera.

The shoulder organ, alias " supraclavicular organ /process " and " postclavicular
organ," lies neither upon nor behind the shoulder-girdle, but within it (Text-fig. 16).

THE "ROSAURA" EXPEDITION 199

Overlooked by nineteenth century authors and their artists (as well as several later
ones) it seems first to have been noted by Brauer (1906) :

" Dorsal von der Brustnosse liegt eine grosser e nach hinten gerichtete
Papille."

Beebe (1933) notes its gross structure, provides the figure here copied, and
mentions ontogenetic changes :

" In cleared specimens the supraclavicular process is seen to be a retort-
shaped sac, with a posteriorly directed neck terminating in a single pore.
This seems to be the only opening, external or internal, to the organ, which
arises just interior to the pectoral girdle at the junction of the supracleithrum
and cleithrum, well below the lateral line. The posterior part of the sac and
the entire neck are exterior to the body wall. It is possible that, instead of
being greatly reduced in relative size in larger specimens, as is generally believed,
the organ merely has a smaller part of its surface projecting exteriorly."

To this account I can add that in the smaller " Rosaura " specimen now available
(Text-fig. 17) there is evident a complication in the neck of the shoulder-organ.
Its opening is oblique, like that of a hypodermic needle, and its duct is supported
by a tubular reinforcement, a modified scale or thin bone (? rudimentary post-
cleithrum). This skeletal structure is flexible, and penetrates deeply towards the
sac of the shoulder organ. Its presence in small Searsids in which the squamation
is undeveloped rather points to its being a specialized bone. I have not proceeded
further with this investigation for the present, since most of the British Museum
material is taxonomically significant as well as scanty, and haphazard dissection
undesirable.

Parr (1937) first attached systematic importance to the shoulder organ in erecting
the genus Searsia. In his latest publication (1951) he omits mention of reinforce-
ment of the external tube, but adds certain information to Beebe's earlier account :

:< The sac is lined with black integument and traversed from outer to inner
wall by irregular strands or columns of soft tissues. There can be no doubt
that this represents a secretory organ, and it seems quite likely that its function
may be to secrete a luminous mucus. The sac discharges through a tube,
also lined with black tissue . . . (this tube) is apparently quite strong
since it seems to withstand considerable abrasion and the loss of all scales."

Is it not possible to go further than this, on the anatomical evidence now available ?
Everything points to the shoulder organ being a structure for the reception of
luminous particles which are effectively blacked-out by the densest pigment in the body
until required, and then expelled in a jet to form a luminous protective cloud, as in the
squid Heteroteuthis dispar and in several deep-sea Crustacea — Acanthephyra pellucida,
Heterocarpus alfonsi, Pandalus alcocki, Aristaeus coruscans, etc. (For encyclopaedic
bibliography see Newton Harvey, 1952, Bioluminescence. New York). It is true
that a ductus ejaculatorius or muscles in the wall of the sac have yet to be demon-

200 THE "ROSAURA" EXPEDITION

strated, but even in the absence of these a mesiad movement of the cleithrum
could readily induce ejaculation : very light pressure on the pectoral girdle of a
preserved specimen produces a strong current of preservative fluid.

If this case is established it is the first instance of a luminous cloud-thrower to
be found among fishes. Ventral saccular luminous organs are known in many
Macrourids and in Physiculus among the Gadids (references in Newton Harvey,
ante), but anatomical evidence and field observation in these cases point to incom-
plete concealment of the luminous contents of the glands, and to a gradual emission
of luminous slime onto the body.

(2) Light organs

Light organs in animals are known or believed to subserve a variety of functions :
attraction of prey ; diversion and confusion of pursuing predators ; local environ-
ment illumination as an aid to vision ; neuro-facilitation of the optic nerve ; self-
indication of the state of the whole photophore system by means of " pilot lights " ;
intra-group (species) recognition ; sex recognition, advertisement and maturity
indication. So much is generally known, but it does not seem to be as generally
realized that the relative evolutionary potentialities and consequent taxonomic
significance of each type of photophore system must vary very widely in relation
to the function performed.

In many cases selection-pressure will favour lighting systems having greater
efficiency, and consequently bring about the convergent evolution of similar arrange-
ments in unrelated groups. Analogous luminous cloud-projection devices in
squids, Crustacea and fishes have been discussed in the previous section (p. 199) ;
special modification of homologous structures may also be replicated, as for example
the terminal photophores on the first dorsal rays of Chauliodus and the Ceratioid
Angler-fishes and the minute light organs on the eyeballs of Chauliodus, Per spar sia
and certain new Myctophids in the " Rosaura " Collection.

Once a luminous organ or organ-complex has attained a certain measure of
efficiency, selection-pressure may, while maintaining that standard, permit such
variation in detail as does not detract from the usefulness of the whole. Thus it
happens that in the illicium (esca) of the Ceratioids and in the mental barbel of the
higher Stomiatoids there is considerable inter-specific (and, to an unknown extent,
intra-specific) variation. In the latter group it has sometimes been possible, without
affecting the attractiveness of the mental barbel to pre}^, to superimpose a sexual
dimorphism and thus fulfil two functions with the same structure.

Only in the case of recognition marks is there a strong tendency towards differen-
tiation of photophore patterns, and that in definite steps with a rigorous disciplining
of the range of variation at each stage. Selection-pressure in favour of such
differentiation is least intense under conditions of allopatric speciation (which in
the marine habitat has great possibilities in the third dimension, e.g., bathymetric
stratification of species of Cyclothone) and strongest under conditions of sympatric
speciation. The tendency is well demonstrated by the wide divergences between
the fundamental photophore patterns of the deep-sea squids, the several families

THE "ROSAURA" EXPEDITION 201

of Stomiatoid fishes and the Myctophids, and by the wonderful diversity within
each of these groups.

It follows, therefore, that in cases where there is reason to believe that intra-
species recognition is the primary function of a photophore system the variations
of that system may be expected to provide the most sensitive index of speciation.
Moreover, variations in recognition patterns, like variations in physiological
tolerances, breeding behaviour, etc., themselves constitute isolating factors, and
may thus precede or even initiate further genetic and therefore morphological
differentiation of populations.

Conversely, if, after making due allowance for the disparity between the taxono-
mist's and the animals' assessment of a variation, that variation does not seem
likely to be clearly characteristic at a distance, then one should be reluctant to
apply it as an index of speciation in sympatric populations.

Having stated these general principles we may now consider the light organs in
the two species introduced by Parr (1937) as Searsia koefoedi and 5. polycoeca.
Modest and probably relatively inconspicuous as photophore systems go, clearly
distinguishable from the basic patterns of the Myctophidae and Stomiatoidea by
their very economy, localized at a few points along the ventral surface, there can
be little doubt that their primary function is intra-specinc recognition. Yet in
these two species the fundamental patterns are, in lateral view, absolutely identical
(assuming no difference of colour) : five lights placed at equivalent points along a
straight line. Only in ventral view does the relatively subtle difference in disposition
and relative sizes of the subventrals and postventrals appear (cf. Parr's figs. 4 and
5), and although the eyes of Searsidae have probably a wide field of vision (particu-
larly in the young stages ; my Text-fig. 19), there is no such obvious connection
between upward binocular vision and downwardly directed photophores, as can
be demonstrated in the Sternoptychidae. The differences between the lateral
views of the photophore systems in these two species and in Persparsia, another
genus of the same family (post, p. 209) support the case for lateral distant recognition.

Beebe, (1926, The Arcturus Adventure, N.Y. : 216, pi. 5) has very beautifully
illustrated the sexual dimorphism in Myctophum coccoi ; the male has conspicuous
luminous scales along the upper side of the tail and the female along the lower
side. To me the divergence between Searsia koefoedi and 5. polycoeca for long
seemed of much the same order and likely significance ; a divergence of far less
use than that in Myctophum as a means of recognition at a distance, but of likely
utility in identifying the sexes to one another in a group of individuals of one species
brought together by the general species recognition pattern of the remaining photo-
phores. On that hypothesis I made a considerable effort (Table II) to justify the
treatment of 5. polycoeca as a synonym of 5. koefoedi, but after the examination of
material of each I am driven to accept them as distinct species. It will be interesting
to see whether the promised ' Dana ' Report on the Searsidae will fulfil the prediction
now made that the Atlantic populations of these two species will prove to be geo-
graphically and /or bathymetrically isolated.

Although this hypothesis of sexual dimorphism in Searsia has proved unfounded,
the general similarity of the photophore patterns remains a powerful argument

202 THE "ROSAURA" EXPEDITION

against the removal of S. polycoeca to another genus — Holtbyrnia — and within
that genus to a sub-genus Mentodus containing one other species, H. (M.) rosiratus
(Giinther), said to be known only from a unique holotype devoid of photophores.

The last-mentioned specimen has been badly cut about by successive investigators,
but there is no doubt that it originally possessed no trace of light organs. The
ventral squamation is complete and intact except at those points where exploratory
borings have been made, and experiments on specimens of the more plentiful
Persparsia taaningi indicate that abrasion of the photophores without damage to
the body is highly unlikely. All the specimens identified as " Bathytroctes rosiratus
Giinther " by other authors and examined by Parr have proved to be either S.
koefoedi or P. taaningi.

Parr (1937) supports his argument against ontogenetic atrophy of the light
organs in B. rosiratus by citing from the literature a series of Searsids of ascending
size in which those structures are present, including one (unfortunately not figured
by its original author) of equivalent size to Giinther 's holotype. The critical
specimen is still 10 mm. S.L. shorter than Gunther's, and this may, in terms of age,
be rather more significant in fishes apparently near their limit and presumably
growing at a correspondingly slower rate. However, Krefft (1953, see my footnote,
p. 206) in a paper received while the present was in the press, describes a photo-
phore system constant in number and arrangement through a series from 87-174
mm. S.L.

A specimen in our collections from " Discovery " St. 2074 (21.ix.37. io° 10' N.,
21 13' W. ; closing net, oblique haul ; 875-400 metres, sounding 5,148 metres)
bears a label " Holtbyrnia polycoeca Id. A. E. Parr." The salient characters of this
fish are given in Table II, where it will be noted that it gives good agreement with
the holotype of that species in all characters except the head length, which runs
contrary to the expected allometry. There are, however, discrepancies in the
light organs. The supraventrals and supra-anals are present and typical, but the
posterior anals are marked only by small patches of dark pigment, the median
subcaudal is very faint and has only a trace of pigment, the thoracic organ is very
thin if it is present at all, and the subventrals and postventrals are altogether
wanting. Since the specimen is in excellent condition I am inclined to regard the
photophore system as undergoing degeneration.

(3) Notes on taxonomy of Searsids

Beebe (1933) has interpreted Gunther's (1887) figure of the snout of Bathytroctes
rosiratus in terms of the typical Searsid condition with forwardly directed tusks,
and this interpretation was at first disputed by Parr (1937). On examining the
specimen I find the premaxillaries thin and somewhat eroded, but the terminal
lip-like projections are still present and prove to be hollow and capable of admitting
a bristle for a fair distance. I regard these structures therefore as representing the
broken stumps of tusks with their related pulp-cavities ; alternatively and less
probably, as the premaxillary sockets of tusks which have been shed (see also
p. 208). The increased number of accessory teeth is a function of age and size,
while the elaborate crenellation of the edge of the jaw is of a type common on the

THE "ROSAURA" EXPEDITION

203

edges of thin bones which have completed their major growth, comparable, for
example, with the elaboration of the skull crests in Melamphdes.

Parr has also complained that Beebe (1933) shows a shoulder organ in a figure
stated to be based on Giinther's when the original and description do not include
it. Perhaps Beebe saw the specimen at some time and erred only in the legend to
his figure, otherwise it must appear very odd that both Beebe and Parr are correct ;
there is a shoulder organ present in the specimen, though not shown in the original
figure. Only the tip of the tube, with its internal support, projects from the body,
but the sac must still be quite large, for a fair amount of preservative fluid and
air bubbles can be extruded from the tube by exerting moderate pressure on the
shoulder girdle.

Parr (1951) modifies his earlier standpoint on this species by placing B. rostratus
fairly among the Searsidae as Holtbyrnia (Mentodus) rostratus. Though he does
not elaborate his reasons in what, after all, only purports to be a preliminary re-
vision, we can fairly assume that he has made the above observations during one
of his vists to the British Museum (Natural History).

The principal characters of three species of Searsids are tabulated in Table II.
A number of identical or nearly identical features are omitted for brevity's sake,
but all the salient points have been included.

Principal specific differences relate to the light organs (already discussed p. 201),
the proportions of the head and upper jaw to standard length, the gill-raker counts,
presence or absence of a dermal pit behind the shoulder, and the number and nature
of the pyloric caeca. Having tested and rejected the hypothesis that 5. koefoedi
and 5. polycoeca were sexes of the same species, I hoped at least to find support
for a theory that 5. polycoeca was a young stage of B. rostratus. Unfortunately
this idea must be rejected also, although I am satisfied that B. rostratus and 5.
polycoeca are very closely related, and confidently expect that the young stages of B.
rostratus when found will prove to possess photophores. I would advocate the
inclusion of all three of these species in the genus Searsia as the expression of this
close relationship.

The importance to be attached to the pyloric caeca may be questioned. Con-
sidering Parr's figures of the pyloric caeca in 5. koefoedi and 5. polycoeca (1937,
figs, ia & b), it may be seen that there is a fundamental similarity between them,
for both have five primary caeca along the right side of the stomach and a further
group of three along the anterior end adjoining these. Even the condition in S.
rostratus which has been described as having " 14 simple, straight pyloric ap-
pendages " can be reduced to this same basic pattern. Although the immediate
impression is of fourteen simple caeca (n + 3) packed side by side like the fingers
of some polydactyl hand, closer inspection shows that there is one group of five
primary caeca, of which the anterior one forks twice, the second, third and fourth
once each and the last is simple, like the remaining group of three. The forking
takes place very close to the stomach, and the sum of the terminal diverticula,
1. 1. 1.4.2.2.2. 1, gives the total of 14. It may fairly be asked how much variation
is due to individual differences and to ontogenetic changes ? The new specimen of
S. polycoeca presents a much simpler picture than the smaller holotype, while the

204

THE "ROSAURA" EXPEDITION

<t> t^ ^

_<S

„

V
Q

. ro^

„

;

M N co

: n

•**

I*"H

•^

.2 CO o o ^

8 .525^

Pi ro^

: N N

C/5

„ .

• <— i

8 ™ fc ^

73
CD
O

<H-I

o
o

o> oo v

: m ^

C/5

•

•a.

"to

c
"3

*X> O CO P*
d cm v

rj- u~> C* N O N N O OOO

f^oo

O "*-oo Tito in O NO N <
NO O h O^^'tinN 1 '!* ..m

CJ M N CJ MVOlOMM I

w M

Qpn'

M\ONa.roO"1-Oi-iOfO ++^
VmOmoooOOOO^m 2 ^ '-T

mO m (JiONNO O N00

m o>ow no ooo o n-^

«no o o o o o

rfOioNootoovoO'^- <!^ >

fOMMM MOlOMM M rr >

MNNOrv.OCONO^Or}- <3 1-~
t^

(nOOUOOOHNta^OlO ■- M

ro MN MVOWJMM o^~

W Ph Ph

—

J«

oc w

m -O o

M 3 4->

en
M

<j=;

en
O

03 -d

(ti

a rt

rrt

o
S

d m

.2 ^
■2 n>^

d ro

d Sv c

3
O

o S

525 £

d d

0> 4)

0) Si

o
'o

co io

o o

M O

„ „

4->

H->

o in

c^O

M co

■^ o

o 00

<U CD
■°^

tn SO

"cd 3

!■§

CD w

i^H

<U i

.5 t"

— a

rt fli

+5 O .w

THE "ROSAURA" EXPEDITION

205

.2 .2 ^2 rf a

-d -a

(D 03 03

<« £ «

03 Ph

03 .03

^".t-i ^, rt

Ph 03

3 o

03 Cm

03 03

£ I

a §
o 51

iS a
a «

03 C

5 ^

£ §

03
3

3 !h

PhX>

3 3

03 03

2 W fcC'^'cl ""

So"^ £j oi fl 13
a £ JS 5 03 Td

O
Ph

03 ^ S^

oj 2 ?5"(3

vh 03 y ■+■>

CL|HJ ,P 03

H M

M O

N N

13 _,

j_l Ah !-, 03 3 _. O

M r-< O t>i

o3 +J o3 m -S g £ ro

pUsl^fl r

^ — -

& a

, 03

03 4->

.ft

i 3

•0

03
03

^3 03

a
3

+ a

" a 03 r-J

^ o a a

a 03 b 03

S.2 c3 >

03 ia i_j ••-I

03
03 O

73 > a

•s 2 -s s ^ ~ '-«

<o n 2 £ d o3 1

3 03 o 03

03 03 tn

O 0..

T3 ^

6 *

03 g

a °

03 ±>

Ph a

O 03

a '-3

M O

H >%-a ^ j
Siva's

» o3 T3

a-fi

p-i a o

3-a.e

'<a -

B4i a

£ § S 3 M S 8

In 03 X +- 1 , " — ' 03

at! -5 !> fa a -§

3 co O,«o X

S P J3 ^

m "n" o

03 03

Li m

1:3 2

mo .a

±J o3

PhCJ

00 03

o3
O

>,a a rt g_

.<a

o 2

io_j3

^j 03

1 ?

2" ^ 03 M

rL-a 03 10
ga^cf^
-Sav:

"*■ bo <^ b

-•of&g

^'a^

tH . H ™ Cl3

S ^1 8

•S m O rt

!_. 03 o a

Oh^ 43 2

hh 03 to a

03 Q

&
%

o {3

03 B3

•° 'H

,a a

-s^ a

03 03 03
4-J 3 M
+J Ph

^a-a

-rr o3

•ii- in

03 r3

>,^ a ^3

03 4_i
03 03
+> 03
03^

a 03

03 +3

a ^

o "a

>Ph

206 THE "ROSAURA" EXPEDITION

larger new S. koefoedi is markedly different from the holotype of that species. It
may well be that the primary forks remain as they are, while the total of tertiary
divercula becomes reduced by subsequent growth at the secondary forks.

The British Museum (Natural History) now possesses the following material of
this family :

i Sear si a rostratus (Giinther) Holotype.

2 Searsia koefoedi Parr.

i Searsia polycoeca Parr.

5 Per spar sia taaningi Parr.

i Normichthys operosa Parr.

I Pellisolus facilis Parr.

As regards Parr's draft classification of 1951, I would modify it as already indi-
cated to show the supposed affinities of the three species which I place in Searsia.
There arises another problem, common enough when reticulate evolution has to
be expressed in a two-dimensional classification, of indicating the affinity which ap-
pears to exist between Normichthys operosa and S. polycoeca alone among the
Searsia group. Normichthys has a row of six large pits behind the shoulder, each
containing three opaque white bodies (? photophores or sense-organs ?), as well
as a smaller pore in every scale of the body. The single dermal pit behind the
shoulder of S. polycoeca seems to be the precursor of such a condition.

The grouping adopted in the key according to relative size of head and length
of upper jaw is useful enough for taxonomic purposes, but does not impress me as a
natural phyletic classification, for the reasons which have already been indicated.

Searsia koefoedi Parr 1

(Text-fig. 17)

Parr, 1937, Bull. Bingham Oceanogr. Coll. 3 : 16, Fig. 4 ; Maul, 1948, Bol. Mus. Municipal
Funchal, No. 3, Art. 5 : 12, Fig. 3 ; Parr, 1951, American Mus. Novitates, No. 1531 : 17.

Bathytroctes (Talismania) homopterus (non Vaillant) (part) Norman, 1930, " Discovery " Reps.
2 : 269 (not fig.).

St. 46. 30.xii.37. 7 27' N., 23 08' W. 2-metre Stramin Net, iooo(-o) metres.
1 specimen, 15-5 mm. S.L. Reg. No. 1953. 3.6. 12.

Hab. North and South Atlantic ; Caribbean.

A young fish and contorted, its body bent downwards from behind the nape
and its tail curved forwards along the left flank. The following information expands
and supplements that given in Table II.

D.20 + ; A.17 + ; P.20±; V.7; C.7 + :XXII:5 +

Branchiostegal rays 7. Gill-rakers on first arch 0/1/9.

Proportions in per cent of length without caudal (S.L.) : length of head, 34-1 ;
snout, 8-3 ; orbit, n-6 ; interocular width, 7-0 ; upper jaw, 167 ; premaxillary,
4-5 ; posterior supramaxillary, 5-8 ; lower jaw, 21-9 ; snout to ventrals, 6o-o ;
snout to dorsal, 66-4 ; snout to anal, 70-9 ; greatest depth of body, 18-0 ; least
depth of caudal peduncle, 6-4 ; dorsal fin base, 16-1 ; anal fin base, 12-9.

1 A paper by Krefft (1953, Zool. Anz. 151 : 259-266) received while the present work was in the press,
describes as new Searsia schnakenbecki from Iceland.

THE "ROSAURA" EXPEDITION

207

* *' .;

in
B

a
o

.-';'-■.';>";

:'<*;

..»;*.■*

2 o8 THE "ROSAURA" EXPEDITION

Least depth of caudal peduncle 357 per cent of greatest depth of trunk.

0-0625 S.L. = 0-97 ; 0-0375 S.L. = 0-58 (data for Parr's formulae).

Dentition. Only the extreme tip of the forwardly directed tusk on the pre-
maxillary is yet erupted. One downwardly directed fang at the front of each pre-
maxillary, followed by five slightly smaller teeth which are fairly regularly spaced.
Ten similar teeth on the maxillary, with indications of smaller teeth to be erupted
between and within these. One pair of fangs on vomer, the largest teeth yet de-
veloped. One slender tooth on each palatine. On the dentary a single row of
numerous very fine teeth, with a few (4-5) larger vertically directed ones outside
these and probably of the same series. Of the horizontally directed series two
teeth are erupted at the front end of each dentary and others unerupted may be
seen opaque white in the jaw. A tiny projection below the dentary symphysis.

Eyes. The orbits are slightly longer than high. The eyes project some way
beyond the general contour of the head and must have wide and partly coincident
fields of vision. The lens is pale green, large but slightly smaller than the pupil
as fixed. There is a greyish-white retinal tapetum lucidum, which may be luminous.

Colour. Dorsally light brown, ventrally dusky. Outer integument over
shoulder organ granular dark brown, with darker internal pigment of tube showing
through. A darker triangular patch between the shoulder organ and the pectoral
fin. A dark line along the lateral line. Fin bases pale.

There are no certain indications of light organs. The fins are well advanced,
but the full complement of anterior dorsal rays, procurrent caudal rays, and possibly
of anterior anal rays, is not complete. Gill-rakers moderately long and widely
spaced ; their complement also incomplete. Lateral line pores before ascending
limb of post-temporal.

This specimen agrees very well with Searsia koefoedi in all general proportions.
The relative distances from snout to dorsal and anal differ by 4-3% (" less than
4% ") — a trivial difference against the likely error in measuring a difficult specimen.
Lower counts for teeth, fin-rays and gill-rakers reflect the relatively small size of
the fish, and in some respects show an approach to 5. polycoeca.

Of the three specimens referred to Bathytroctes (Talismania) homopterus Vaillant
by Norman (1930), the largest, from " Discovery " St. 269, proves to be Normichthys
operosa Parr, as already noted in Parr (1951). The second specimen, from the
same station, seems no longer to be in this Museum. The third, from an un-
numbered " Discovery " station, is a fine Searsia koefoedi whose characters are
given in Table II. It is of interest to note that in this specimen the left premaxillary
tusk is laterally duplicated, which may be an indication of periodic replacement
and so provides a possible explanation of the tusk-less condition of the holotype of
Searsia ro stratus.

Persparsia taaningi Parr

(Text-figs. 18 and 19)

Parr, 1951, American Museum Novitates No. 1531 : 18.

Bathytroctes rostratus (nee Gunther) Norman, 1930, " Discovery " Reps. 2 : 268, Fig. 1 ; PI. 2,

Fig. 3 ; Beebe, 1933, Zoologica, N.Y. 16 : 43, Figs. 8c, 8d.
Searsia n. sp. Parr, 1937, Bull. Bingham Oceanogr. Coll. 3 : 16.

THE "ROSAURA" EXPEDITION 209

St. 33. 22.xi.37. ii° 00' N., 75 43' W. ; 2-metre stramin net ; c. i,20o(-o)
metres. 1 specimen, 23 mm. S.L. Reg. No. 1953. 3. 6. 13.
Hab. North and South Atlantic ; Caribbean.

A young fish and badly contorted, the posterior half of the body twisted and
bent forward to lie along the left cheek.

D.21; A. 14; P.20±; V.8; C.3 + : XXII : 2 +-.

Branchiostegal rays 7. Gill-rakers on first arch 4/1 /12.

Proportions in per cent of length without caudal (S.L.) : length of head, 40-0 ;
snout, io-8 ; orbit, 13-9 ; interocular width, 6-o ; upper jaw, 20-8 ; premaxillary,
6-5 ; posterior supramaxillary, 7-8 ; lower jaw, 25-2 ; snout to ventrals, 62-1 ;
snout to dorsal 63-4 ; snout to anal, 71-3 ; greatest depth of body, 16-5 ; least
depth of caudal peduncle, 6-5 ; dorsal fin base, 20-4 ; anal fin base, 12-1.

Least depth of caudal peduncle 39-4 per cent of greatest depth of trunk.

0-0625 S.L. = 1-437 ; 0-0375 S.L. = 0-862 (data for Parr's formulae).

Dentition. The forwardly directed tusks are present but not erupted. The
premaxillaries are very thin and their edges damaged ; teeth that remain at the
hinder ends are like those of the maxillary. The maxillary has about 20 teeth
with indications of very fine ones between these. Four fangs on vomer, the largest
teeth yet developed. Two slender teeth on each palatine. The dentary has a
small upwardly directed projection at the symphysis and many acicular teeth in a
single row.

Eyes. The eyes and orbits are elongated. The eyes project some way beyond
the general contour of the head, and must have wide and partly coincident fields
of vision, especially in the upward direction (Text-fig. 19). The lens is pale green
and as large as the pupil as fixed. There is a greyish-white retinal tapetum lucidum.

Colour. Dorsally light brown, ventrally dusky. Integument over shoulder
organ dark brown. A dark line of pigment along the lateral line. Fin bases pale.

There are no certain indications of light organs. Gill-rakers long, the longest
0-9 mm. (3-9% S.L.). Gill-filaments numerous, short and stumpy. No lateral
line pores before ascending limb of post-temporal.

This fish shows a number of small divergences from Parr's diagnosis (1951),
especially in the slightly longer head, longer lower jaw, and more posterior insertion
of the ventrals. In view of the youth and twisted condition of the specimen it
seems better not to attach much importance to these differences in face of the
general agreement with P. taaningi.

Norman's (1930) figures from " Discovery " specimens which I identify with
this species do not do full justice to the photophore system. In addition to those
shown in his drawings and plate there are : two light organs on the eyeball, 1 in the
upper and lower anterior corners ; one on the skin over the lower posterior edge
of the orbit ; one near the hinder edge of the operculum, above the level of the
pectoral fin base ; in the largest specimen (47 mm. S.L.) a median one at the lower
ends of the cleithra, and in the same specimen an additional median photophore
between the supraventrals (not to be confused with the curious two-way light

1 Searsia schnakenbecki Krefft (1953) has a photophore on the eyeball. Parr (1937) seems to figure one
in 5. koefoedi though he makes no mention of it in his text.

THE "ROSAURA" EXPEDITION

_ 6

ho
a
s
o

o
.d

THE "ROSAURA" EXPEDITION 211

organ already noted by Norman). In addition to these special organs there are
likely indications of a strip of luminous tissue extending from the occiput to the
dorsal, and elsewhere about the fin bases. The few specimens available are small,
and, while they provide indications that the light organs are not developed simul-
taneously, they are inadequate for a study of possible sexual dimorphism. The
light organs on the eyeball are quite inconspicuous in the largest specimen avail-
able, but there is otherwise no evidence to support ontogenetic atrophy ; on the
contrary. Doubtless Parr will be dealing with this matter in full in his forthcoming
monograph.

Family CLUPEIDAE

Pellonula afzeliusi Johnels

Johnels, 1954, Arkiv. f. zool, Stockholm (2) 6 : 352.

Pellonula vorax (non Gunther) Svensson, 1933, K. Svensk. Vetensk. Handl. (3) 12, No. 3 ; 47, fig.
16 (young stages).

St. 48. 3.1.38. 13 27' N., 15 47' W. ; 2-metre stramin net, 3(-o) metres.
526 specimens, 4-81 mm. S.L. Reg. No's. 1953.3.6. 14-541.

Hab. Gambia River.

Apart from the other characters noted by Johnels, this species differs from P. vorax
in the possession of much more trenchant keeled scutes between the isthmus and the
ventrals.

Harengula pensacolae Goode & Bean

Clupea humeralis (part) Gunther, 1868, Cat. Fish. Brii. Mus., 7 : 422 ;

Harengula pensacolae Goode & Bean, 1879, Proc. U.S. Nat. Mus. 2 : 152 ; Regan, 1917, Ann.
Mag. nat. Hist., (8) 19 : 389.

St. ( — ) 28.x. 37. 17 28' N., 88° 11' W. ; dip-net, at surface. 5 specimens,
60-98 mm. S.L. Reg. No's. 1953.3.6.542-6.
Hab. Gulf of Mexico ; Caribbean.

Ilisha africana (Bloch)

Clupea africana Bloch, 1795, Naturgesch. Ausldnd. Fische, 9 : 45, pi. 407. (Type loc. Accra,

Gold Coast.)
Clupanodon africanus Lacepede, 1803, Hist. Nat. Poissons, 5 : 469, 471.
Pellona iserti Cuvier & Valenciennes, 1847, Hist. Nat. Poissons, 20 : 307.
Pellona gabonica Dumeril, 1858, Arch. Mus. Paris, 10 : 259, pi. 23, figs. 3, 3a.
Pellona africana Bleeker, 1863, Natur. Verh. Holl. Maatsch. Wetensch. (2) 18 : 122, pi. 27,

fig. 1.
Pristigaster dolloi Boulenger, 1902, Proc. zool. Soc. Lond. : 271, pi. 30, fig. 3.
Ilisha melanota Derscheid, 1924, Rev. Zool. Afr. 12 : 278 ; Irvine, 1947, The Fishes & Fisheries

of the Gold Coast, London : in, fig. 39.
Ilisha dolloi Giltay, 1935, Bull. Mus. roy. Hist. nat. Belg. 11, fasc. 36 : 3, figs. 1 and 2.
Ilisha africana Fowler, 1936, Bull. Amer. Mus. nat. Hist. 70 : 178, fig. 71. Delsman, 1941,

Mem. Mus. roy. Hist. nat. Belg. (2) Fasc. 21 : 50.
Pristigaster martii {non Agassiz) Fowler, 1936, op. cit. : 180, fig. 72.

[non Pellona africana Gunther, 1868, Cat. Fish. Brit. Mus., 7 : 455.]
[non Ilisha africana Norman, 1922, Ann. Mag. nat. Hist. (9) 11 : 4.]

ZOOL. 2, 6. 13

212 THE "ROSAURA" EXPEDITION

St. 48. 3.1.38. 13 27' N., 15 47' W. 2-metre stramin net, 3(-o) metres.
2 specimens, 84-102 mm. S.L. Reg. No's. 1953.3.6.547-8.

Hab. Coastal waters and estuaries of W. Africa, Senegal to Angola.

Norman (1922) includes Ilisha africana in the first section of his key to the genus
llisha :

"I. Less than 46 scales in a longitudinal series ; 16-20 scutes before insertion
of pelvics."

Bloch (1795) does not supply this information, but counts from his figure give
46 and 24 + respectively, and therefore rather better agreement with the second
section of Norman's key :

" II. More than 46 scales in a longitudinal series ; 20-28 scutes before
insertion of pelvics."

At the time when Norman wrote, the British Museum (Natural History) did
not, in fact, possess any material of Bloch's species, nor indeed any belonging to
the genus Ilisha and of undoubted African origin. Norman followed Giinther
(1868) in mistakenly identifying with Bloch's fish a single specimen, B.M. (N.H.)
No. 1852.9. 13. 175, which came originally, and without any locality data what-
soever, from the collections of the Zoological Society. This specimen should be
assigned to 7. indica (Swainson) or I. brachysoma (Bleeker), two doubtfully distinct
species from India and the East Indies, from which region it very probably came.
Its ventral scutes (20 + 8) give a lower pre-pelvic count than in any African Ilisha
known, and are much less salient than those in the copious African material of 7.
africana (Bloch), (including specimens from Accra, the type locality), received at
this Museum in recent years.

Norman's error has effectively bedevilled every subsequent author except Fowler
(1936), who providentially overlooked his paper in the main body of The Fishes of
West Africa and later mentions it without comment in an " Appendix " to this
work.

Derscheid (1924) leaned heavily upon Norman's revision and advice and (Bloch's
book being extremely rare) seems not to have checked the original description and
figure of I. africana. Taking very literally Norman's statement that 7. africana
(non Bloch) is " scarcely distinct from 7. brachysoma," Derscheid has apparently
deduced a pseudo-description of the genuine 7. africana by so modifying Norman's
detailed account of 7. brachysoma as to reconcile it with the few minor differences
noted by Norman in his brief comments on a supposed I. africana which was, as
we have seen, not Bloch's species. Derscheid then proceeds to detail the differences
between his African material from the mouth of the Congo and this pseudo-
I. africana whose antecedents, if they may be said to have any reality at all, are
Asiatic. Naturally enough he fails to recognize his own material as the genuine
7. africana (Bloch) and so 7. melanota joins the lengthening list of synonyms. Ex
Africa semper aliquid novum I

Giltay (1935), having examined the types of 7. melanota Derscheid and of Pristi-
gaster dolloi Boulenger (1902), makes 7. melanota a synonym of 7. dolloi (Boulenger).

THE "ROSAURA" EXPEDITION 213

This view will be supported in greater detail presently ; for the moment it is merely
emphasized that Giltay did not carry his synonymy back any further than Boulenger
(1902), and merely includes " Ilisha africana Norman . . . (pro parte) " as an
allusion to Norman's tentative synonymy of P. dottoi Boulenger with /. africana
(Bloch).

Fowler (1936) overlooked Derscheid's paper as well as Norman's, and conse-
quently makes no contribution to the brachysoma-africana-melanota tangle. In-
directly, however, he adds to the confusion by making P. dolloi Boulenger a synonym
of Pristigaster martii Agassiz (Spix & Agassiz, 1829, Pise. Brazil : 55, pi. 24a) and
presenting a re-drawn version of Boulenger's figure of 1902 as " Pristigaster martii,
modified from Boulenger." Ludicrous as it may seem after a comparison with the
original figures, the error is liable to prove as dangerous as Norman's, for Spix &
Agassiz is almost as rare a work as Bloch.

Delsman (1941) has questioned the validity of /. melanota, but upon the same
false premises as led Derscheid to erect his species. Delsman concentrates upon
the ventral scutes, which Derscheid gives as 18-20 + j-8 in his pseudo-L africana
based on Norman's account of /. brachysoma and 25-27 + y-8 in /. melanota, and
endeavours to close the imaginary gap by citing counts of 22-24 + 7 from new
material. Bloch's original figure (1795) shows ca. 24 + 5 ; if there is any problem
it is presented by the post-pelvic rather than the pre-pelvic count, and if we allow
the artist a small error in the pelvic fin insertion we can reconcile this difference
without compromising the opening argument in the present article, for it is possible
to postulate a correction from ca. 24 + 5 to ca. 22 + 7 without even invoking the
probable scutes concealed by the operculum in Bloch's figure.

Norman (1924), Giltay (1935) and Delsman (1941) have all, from their several
concepts of /. africana, queried the distinctness of Pristigaster dolloi Boulenger
(1902). Boulenger introduced this species with a brief description and figure, and
no discussion. Derscheid (1924), having compared the type of P. dolloi with his
types of /. melanota, is unable to discover any other important difference than
that the former lacks even the rudiments of pelvic fins. Giltay, working also from
original material, gives an intensive comparison, together with a theoretical disserta-
tion on the body-forms of fishes which, curiously enough, fails to make the obvious
comparison with the Gastropelecidae. Giltay's very valuable contribution is to
cite the work of Pellegrin (1926, C.R. Acad. Sc. Paris, 183 : 1301) who examined
13 specimens of the African Clariid Clariallabes variabilis Plgr. and found 6 with
both ventrals present, 2 with one only, and 5 with none. Hildebrand (1943, Bull.
Bingham Oceanogr. Coll. 3 Art. 2 : 140) reports 1 specimen lacking ventrals among
20 Pterengraulis atherinoides (L.) examined by him. Considering such precedents,
the minute pelvic fins in Ilisha, and the close correspondence between Boulenger's
unique holotype of P. dolloi and the many African specimens of Ilisha now examined
by several authors, there exist no reasonable grounds for retaining Boulenger's
species as distinct.

The considerable African material of I. africana (Bloch) received at the British
Museum since Norman's time may properly be referred to Ilisha Gray & Richardson
as generally understood, since it possesses a maxillary not adherent to the pre-

2I 4

THE "ROSAURA" EXPEDITION

maxillary, and rounded at the posterior end ; a ligament rather than a bone between
the maxillary and premaxillary (but see Myers, Copeia, 1950 : 63), and no canine
teeth.

A new world-revision of Ilisha, desirable as it appears, cannot be undertaken in
the middle of a general report. I state therefore only a general impression that,
on the material which I have handled for comparative purposes, there may be a
case for recognizing the distinctness of Ilisha africana from the Indo-Pacific forms
by the creation of a new sub-genus. Compared with /. br achy soma, I. indica,
I. kampeni, I. sladeni, etc., it shows the following differences : the ventral scutes
are far more salient ; the lower jaw projects further, so that the mouth opening
is carried further dorsad ; the lower part of the body (ventral to the lower jaw) is
more developed and keel-like, so that the overall body depth is greater and the
pectoral fin, though maintaining the same relations to the head, is inserted pro-
portionally higher ; the pectoral fin, though obliquely inserted, tends to twist so
as to project in a horizontal plane perpendicular to the surface of the body. In
these same characters it approaches the related genera and species of the Atlantic
coasts and rivers of South America.

Perhaps the most interesting feature of Ilisha africana is that in *he development
of the deep ventral keel, in the carriage of the pectoral fins, elongation of the anal,
exaggeration of the lower lobe of the caudal, and incipient loss of the pel vies,
(though not in the normal lateral septum), it foreshadows the form and probably
also the habit of the Gastropelecidae — South American Ostariophysi which " fly "
by " a taxi-ing movement, with the large thoracic keel cleaving the water and the
pectoral fins beating the surface " (Fraser-Brunner, 1950, Ann. Mag. nat. Hist. (12)
3 : 959). These characteristics of I. africana attain a full convergent evolution of
the Gasteropelecid condition in the related South American Pristigaster cayanus
Cuvier (1817, Rcgne Animal, 4, pi. 10, fig. 3). (P. martii Agassiz is a synonym of
this species.) Parallel evolution of the same condition has occurred in the Indo-
Pacific Opisthopterus, in which the correspondence to the Gasteropelecid form is
increased by the smooth ventral profile and the more posterior insertion of the
dorsal, which in this genus commences behind the anal origin.

4->

' — -

d
cd

-p

"aT

Ph,

o«

^
+J

Bull. B.M. {N.H.) Zool. 2, 6.

PLATE 8.

Electric Rays, Diplobatis pictvts Palmer. Two female specimens, actual size, from " Rosaura'
Station 36. The outlines have been retouched. (Photo British Museum (Natural History).)

5 . SPONGES

By MAURICE BURTON

With Plate 9 and Text-figs. 1-9

SYNOPSIS

The 42 specimens, representing 35 species, of which 8 are new, were collected from widely-
separated points on the Atlantic seaboards, providing useful fauna records. Two specimens
are of special interest, since they make possible re-descriptions of species long known but too
inadequately described to make recognition possible. The species are Stryphnus pachastrel-
loides (Schmidt) and Callyspongia tenevrima (Duchassaing and Michelotti).

I. LIST OF SPECIES MENTIONED IN THIS REPORT

Order Hexactinellida.

Sub-order Amphidiscophora.

Family HYALONEMATIDAE.

Pheronema carpenteri (Thomson).

Sub-order Hexasterophora.

Family FARREIDAE.

Farrea occa Bowerbank.

Family MELETTIONIDAE.

Aphrocallistes beatrix Gray.

Family ROSSELLIDAE.

Ros sella mortenseni Burton.

Order Calcarea.

Family HOMOCOELIDAE.

Leucosolenia botryoides (Ellis and Solander).
Leucosolenia canariensis (Michlucho-Maclay) .

Order Tetraxonida.

Sub-order STREPTASTROSCLEROPHORA.
Family THENEIDAE.

Thenea fenestrata (Schmidt) .
Thenea wyvilli Sollas.

216 THE "ROSAURA" EXPEDITION

Sub-order Astrosclerophora.

Family STELLETTIDAE.

Stelletta anancora (Sollas).
Stryphnus pachastrelloides (Schmidt) .

Family CLAVULIDAE.

Radiella sol Schmidt.
Family TETILLIDAE.

Tetilla cranium (Miiller).

Sub-order Sigmatosclerophora.
Family HAPLOSCLERIDAE.

Haliclona rubens (Pallas) .

,, erina de Laubenfels.

,, spiculosa (Dendy).

,, calcinea sp. n.

,, tenerrima sp. n.
Callyspongia tenerrima Duchassaing and Michelotti.

Family DESMACIDONIDAE.

Desmacella annexa (Schmidt) .

inornata (Bowerbank).
Myxilla distorta sp. n.

Phorbas amaranthus (Duchassaing and Michelotti).
Inflatella viridis (Topsent).
Plocamionida topsenti sp. n.
Tedania anhelans (Lieberkiihn) .

Family Axinellidae.

Axinella ramosa sp. n.
Hymeniacidon assimilis (Levinsen).

,, glabrata sp. n.

Halichondria panicea (Pallas).

,, osculum Lundbeck.

,, bowerbanki Burton.

,, cornuloides sp. n.

Family RASPELIIDAE.

Higginsia strigilata (Lamarck) . ,
Dragmatyle topsenti sp. n.

Order Keratosa.

Family SPONGIIDAE.

Hircinia variabilis Schmidt.

THE "ROSAURA" EXPEDITION 217

II. SYSTEMATIC NOTES

Order Hexactinellida.

Sub-order Amphidiscophora.

Family HYALONEMATIDAE.

Genus Pheronema Leidy.

Pheronema carpenteri (Thomson).

Holtenia carpenteri Thomson, 1869, p. 120, woodcut ; Thomson, 1869, p. 210 ; Thomson,

1869, p. 702, pis. lxvii-lxxi ; Pheronema carpenteri, Kent, 1870, p. 243, pi. lxiii, fig. 1 ; Gray,

1870, p. 210 ; Holtenia carpenteri, Bocage, 1871, p. 69 ; Pheronema carpenteri, Marshall, 1876,
p. 130 ; Schulze, 1886, p. 64 ; Schulze 1887, p. 241, pi. liii ; Schulze, 1893, p. 562 ; Schulze,
1904, p. 50, pi. xv, figs. 1-5 ; Burton, 1928, p. 14.

Occurrence. St. 49, between Fuerteventura Island (Canaries) and Africa
(28 25' N., 13 34' W.), 1st February, 1938. One specimen.

Distribution. Iceland ; Faroes ; Lewis ; Portugal ; Brazil ; Zanzibar ; 823-
2928 m., on mud.

Remarks. In spite of the wide geographical range, and the number of references
to this species in the literature, it has not been recorded on more than five occasions.
Yet from the numbers of individuals referred to in those references the species would
appear to be abundant at suitable depths. This is borne out by the account Bocage
(I.e.) gives. He refers to it as well-known to the " pecheurs des squales " in the deep
waters off Se tubal.

In dealing with this species in 1928, 1 gave a reference to " Carpenter and Thomson
1869 " both in the synonymy list and in the list of literature. This was taken from
Lendenfeld's bibliography in his work of 1886 (p. 598). There are two papers listed
by Lendenfeld under this supposed joint authorship, both for the year 1869, and
both were written, in fact, by Thomson alone.

Another correction is necessary, this time from Marshal (1876). This author
lists the genus " Holtenia Schmidt " on p. 126, for H. pourtalesii and, on p. 130,
gives " Holtenia Schmidt p.p. " under Pheronema. The only valid authorship of
Holtenia is, of course, of Thomson, 1869.

Sub-order Hexasterophora

Family FARREIDAE

Genus Farrea Bowerbank

Farrea occa Bowerbank
(for discussion see Ijima, 1927)

Occurrence. St. 49, between Fuerteventura Island (Canaries) and Africa
(28 25' N., 13 34' W.), 1st February, 1938, 1300 m. One specimen.
Distribution. World-wide, from 204-1901 m.

218 THE "ROSAURA" EXPEDITION

Remarks. I find it difficult to believe that the many subspecies, into which Ijima
divided this species, represent anything more than simple fluctuating variations, or,
at most, ecological varieties.

Family MELETTIONIDAE

Genus Aphrocallistes Gray

Aphrocallistes beatrix Gray

Aphrocallistes beatrix Gray, 1858, p. 114, pi. xi ; Ijima, 1927, p. 286, pi. xxiv, figs. 20-30,
pi. xxv, figs. 1-25, text-fig. 35.

Occurrence. St. 34, off St. George, Grenada, West Indies (12 25' N., 6i°4c/\Y.),
27th November, 1937, 720-800 m. One specimen.

Distribution. South-west of Ireland ; Azores ; Madeira ; Canaries ; Cape
Verde Islands ; Ascension Island ; Bermuda ; Florida ; West Indies ; Indian Ocean
(south of Bombay, Bay of Bengal, Andaman Islands) ; Malay Area ; Philippines ;
Japan ; 105-1633 m., on gravel, sand, mud and ooze.

Family ROSSELLIDAE

Genus Rossella Carter

Rossella mortenseni Burton

Rossella mortenseni Burton, 1928, p. 9, figs. 3-9.

Occurrence. St. 49, between Fuerteventura Island (Canaries) and Africa
(28 25' N., 13 34' W.j, 1st February, 1938, 1300 m. One specimen.
Distribution. Iceland ; Faroes ; 479-957 m.

Order Calcarea

Family HOMOCOELIDAE

Genus Leucosolenia Bowerbank

Leucosolenia botryoides (Ellis & Solander)

Spongia botryoides Ellis and Solander, 1786, p. 190, pi. lviii, figs. 1-4 ; Leucosolenia botryoides,
Arndt, 1939, p. 4, fig. 1.

Occurrence. St. 10, NW. Arm, Horn Bay, Newfoundland (52 13' N., 55
47' W.), 23rd September, 1937, 15 m. One specimen.

Distribution. Arctic ; Atlantic coasts of Europe ; Mediterranean ; West
Africa ; South Africa ; Red Sea ; Australia (west and east coasts) ; Atlantic coast
of North America ; California ; Chile ; Sandwich Isles ; mainly littoral, but
occurring down to 860 m., growing on rocks, stones and seaweeds.

THE "ROSAURA" EXPEDITION 219

Leucosolenia canariensis (Michlucho-Maclay)

Nardoa canariensis Michlucho-Maclay, 1868, p. 230 ; N. sulphurea Michlucho-Maclay, 1868,
p. 230 ; AT", rubra Michlucho-Maclay, 1868, p. 230 ; Tarroma canariense, Haeckel, 1870,
p. 244 ; T. rubrum, Haeckel, 1870, p. 245 ; T. sulphur eum, Haeckel, 1870, p. 245 ; Ascaltis
canariensis, Haeckel, 1872, p. 52, pi. ix, figs. 1-3, pi. x, fig. 1 ; A uloplegma canariensis, Haeckel,
1872, p. 52; Ascuris arredifae Haeckel, 1872, p. 52 ; A. papillata Haeckel, 1872, p. 52;
Ascetta compacta Schuffner, 1877, p. 404, pi. xxv, fig. 9; A. coriacea var., Fristedt, 1885,
p. 8 ; Ascuris canariensis, Lackschewitsch, 1886, p. 338 ; Leucosolenia nausicae Lackschew-
itsch, 1886, p. 300, pi. vii, fig. 1 ; L. nanseni Breitfuss, 1896, p. 427 ; Breitfuss, 1898, p. 106,
pi. xii, figs. 1-9 ; Breitfuss, 1898, p. 13 ; L. coriacea, Arnesen, 1901, p. 10 ; Arnesen, 1901,
p. 67 ; L. canariensis (pars), Thacker, 1908, p. 762, pi. xl. fig. 3, text-figs. 157-160 ; Clathrina
canariensis var. compacta Row, 1909, p. 184 ; Leucosolinia nanseni Breitfuss, 1911, p. 311 ;
L. canariensis, Dendy and Row, 1913, p. 724 ; L. nanseni, Derjugin, 1915, p. 289 ; L. canari-
ensis, Hentschel, 1916, p. 4 ; Hozawa, 1918, p. 528, pi. lxxxiv, fig. 2 ; Breitfuss, 1932,
p. 240 ; Hozawa, 1933, p. 2, pi. i, fig. 1 ; Topsent, 1934, p. 7.

Occurrence. St. 10, NW. Arm, Horn Bay, Newfoundland (52 13' N., 55°
47' W.), 23rd September, 1937, 15 m. Two specimens.

Distribution. Spitzbergen ; Kola Fiord ; Norway ; Sweden ; Mediterranean ;
Canaries ; Cape Verde Islands ; Mauritius ; Suez ; Japan ; on rock, sand, stones,
mud, shells ; littoral to 165 m.

Remarks. The similarity between the external form of this species and that of
the better-known L. coriacea tempts the suggestion that it is a recurrent mutant of
the latter. Both occupy a similar range, have the same colour varieties and a
closely similar external form. If, as we have reason to believe, the presence or
absence of oxea in calcareous sponges has no specific value, then the affinity between
L. canariensis and L. coriacea is even closer, and the recognition of the former as a
mutant of the latter depends mainly on the presence of quadriradiates.

Order Tetraxonida

Sub-order Streptastrosclerophora

Family THENEIDAE

Genus Thenea Gray

Thenea fesestrata (Schmidt)

Tisiphonia fenestrata Schmidt, 1880, p. 71, pi. x, fig. 2 ; Thenea fenestrata, Sollas, 1886, p. 185 ;
T. wrightii Sollas, 1886, p. 185 ; Sollas, 1888, p. 63, pi. viii, figs. 11-20 ; T. fenestrata, Sollas,
1888, p. 71, pi. viii, figs. 1-8 ; Ancorina {Thenea) fenestrata, Lendenfeld, 1903, p. 55 ; A. (T.)
wrightii, Lendenfeld, 1903, p. 58 ; T. fenestrata, Wilson, 1904, p. 88, pi. xiii, figs. 2-4, 6-7, 9 ;
T. echinata Wilson, 1904, p. 91, pi. xii, figs. 1-9 ; T. lamelliformis Wilson, 1904, p. 95,
pi. xii, figs. 10-13, pl- x iii» fig- *•

Occurrence. St. 26, north of Turneffe Island, West Indies (17 53' N., 84
44' W.), 7th November, 1937, 900 m. One specimen.

Distribution. Galapagos ; Chile ; West Indies ; Sierra Leone ; 1745-41 14 m.,
on mud and ooze.

220 THE "ROSAURA" EXPEDITION

Thenea wyvillii Sollas

Thenea wyvillii Sollas, 1886, p. 184 ; Sollas, 1888, p. 74, pi. vi, figs. 1-9 ; T. calyx Thiele, 1898,
p. 24, pi. v, figs. 9-10; Ancorina (Thenea) wyvillii, Lendenfeld, 1903, p. 56 ; A. (T.) grayi
calyx, Lendenfeld, 1903, p. 57.

Occurrence. St. 49, between Fuerteventura Island and Africa (28 25' N.,
I 3° 34' W.), 1st February, 1938, 1300 m. One specimen.

Distribution. Off Zebu, Philippines, 174 m., on blue mud.

Remarks. The species evidently ranges throughout tropical waters. In addition
to the present specimen from the tropical Atlantic and the holotype from the
Philippines, there is a third specimen in the John Murray Collection, in the British
Museum, from the Gulf of Aden.

Sub-order Astrosclerophora

Family STELLETTIDAE

Genus Stelletta Schmidt

Stelletta anancora (Sollas)

Pilochrota anancora Sollas, 1886, p. 189 ; P. gigas Sollas, 1886, p. 190 ; P. tenuispicula Sollas,
1886, p. 190 ; P. crassispicula Sollas, 1886, p. 190 ; P. gigas Sollas, 1888, p. 124, pi. xx, figs.
1-13 ; P. tenuispicula Sollas, 1888, p. 127, pi. xv, figs. 28-32 ; P. crassispicula Sollas, 1888,
p. 128, pi. xiv, figs. 9-15 ; P. anancora Sollas, 1888, p. 132, pi. xiv, figs. 16-22 ; Stelletta crassi-
spicula, Lendenfeld, 1903, p. 42 ; 5. tenuispicula, Lendenfeld, 1903, p. 42 ; S. anancora, Lenden-
feld, 1903, p. 43 ; 5. gigas, Lendenfeld, 1903, p. 43 ; 5. crassiclada Lendenfeld, 1906, p. 281,
pi. xxxi, figs. 3-12 ; 5. crassispicula, Topsent, 1922, p. 1 ; Topsent, 1928, p. 123, pi. i, fig. 25,
pi. v, fig. 13.

Occurrence. St. 44, St. Paul's Rock, 28th December, 1937, 50-60 m. One
specimen.

Distribution. Bermuda ; Cape Verde Islands ; St. Paul's Rock ; Bahia ; on
rock and calcareous algae ; 12-91 m.

Remarks. Topsent (1922) has given an account of the variations in this species,
and on the basis of this has suggested the identity of 5. crassiclada and S. crassi-
spicula. It is difficult to understand why he failed at the same time to recognize
Sollas' Pilochrota anancora, P. gigas and P. tenuispicula as synonyms also ; the
species name is, by priority, Stelletta anancora.

Genus Stryphnus Sollas

Stryphnus pachastrelloides (Schmidt)

(Text-fig. 1)

Ancorina pachastrelloides Schmidt, 1870, p. 68 ; Sollas, 1888, p. 205 ; Lendenfeld, 1903, p. 66.

Occurrence. St. 49, between Fuerteventura Island and Africa (28 25' N.,
x 3° 34' W.), 1300 m., 1st February, 1938. One specimen.

THE "ROSAURA" EXPEDITION

221

Remarks. Schmidt's original description of this species, alone, was inadequate
for purposes of identification. With the few details he gives, together with a poor
spicule preparation made by him and now in the British Museum (Reg. No.
70. 5 .3.48), it is possible to recognize the species when these two things can be
compared with fresh material. The present specimen, although collected in the
eastern Atlantic, whereas the holotype was taken off Florida, is massive, with a

Text-fig i. Stryphnus pachastrelloides (Schmidt), a, oxeote, x 50; b > dichotriaene,
X 50 ; c, microxeote, x 500 ; d, microstrongyle, x 500 ; e, amphiaster, x 500.

222 THE "ROSAURA" EXPEDITION

cylindrical outgrowth. There is one oscule at the summit of the main mass and one
at the end of the outgrowth, each leading into a deep cloaca. The colour, in spirit,
is a pale yellowish-brown. Schmidt described the sponge as " unregelmassige
Knollen mit unregelmassig cylindrischen Fortsatzen." The only other guide
to the external appearance is given : " Die Rindenschicht wird lediglich durch . . .
ein braunliches Pigment vertreten." The skeleton of the " Rosaura " specimen
consists of large radially-arranged oxea, short-shafted dichotriaenes, with microxea,
roughened microstrongyla and amphiasters. There appear to be present, also,
oxyasters, but these, I believe, are no more than the amphiasters seen in end-view.
The spicule-preparation (70.5.3.48) contains these same categories, and they are of
similar dimensions in each case. There are, however, a number of smooth strongyla,
about the same length as the microxea but much thicker. At first sight, the
strongyla seem proper to the sponge but the discovery of a number of other more
obviously extraneous spicules in this preparation suggests that the strongyla may
also be foreign to it. Schmidt's preparations, of which the British Museum has
many, are notorious for either the poverty of their proper spicules or the richness
of those spicules accidentally included. The identity of Schmidt's holotype, based
on B.M. 70.5.3.48, and the present specimen is made the more certain by the
presence, in both, of a faint annulus at the centre of so many of the microxea.

The species may be re-described :

Irregularly massive with cylindrical outgrowths ; surface uneven, hispid ; oscules
large, apical ; texture firm, incompressible ; colour, in spirit, light brown ; skeleton
of radially-arranged oxea and short-shafted dichotriaenes, with microxea, micro-
strongyla and amphiasters for microscleres :

Spicules : Oxea, slightly curved, 3-2 by 0-06 mm., dichotriaenes, cladome
o-6 mm. across, rhabdome 1-5 by 0-048 mm., microxea, 0-14 by 0-004 mm.,
microstrongyla 0-024 by 0-004 mm., amphiasters 0-028 mm. across.

Family CLAVULIDAE

Genus Radiella Schmidt

Radiella sol Schmidt

Trichostemma hemisphaericum Sars, 1869, p. 259 [nom. nud.] ; Radiella sol Schmidt, 1870,
p. 48, pi. iv, fig. 6 ; Trichostemma hemisphaericum Sars, 1872, p. 62, pi. vi, figs. 1-15 ;
Radiella sol, Burton, 1930, p. 510 ; Burton, 1934, p. 15.

Occurrence. St. 26, north of Turneffe Island, West Indies (17 53' N., 87
44' W.), 900 m. One specimen.

Distribution. Greenland ; between Norway and Bear Island ; Barents Sea ;
Kara Sea ; Franz Josef Land ; Norway (Lofoten) ; West Indies ; 97-1167 m., on
clay and mud.

THE "ROSAURA" EXPEDITION 223

Family TETILLIDAE

Genus Tetilla Schmidt

Tetilla cranium (Muller)

Alcyonium cranium Muller, 1776, p. 255 ; Tethya cranium, Lamarck, 1815, p. 71 ; Craniella
cranium, Sollas, 1888, p. 51 ; Tethya cranium, Topsent, 1913, p. 12, pi. iii, fig. 3, pi. v, fig. 12.

Occurrence. St. 34, off St. George, Grenada, West Indies (12 25' N., 6i° 49'),
27th November, 1937, 720-800 m. One specimen.

Distribution. Arctic ; Atlantic coast of Europe ; Azores ; West Indies ;
Ceylon ; Japan ; Pacific coast of North America ; 9-1829 m., on rock, gravel, sand,
mud and ooze.

Family HAPLOSCLERIDAE

Genus Haliclona Grant

Haliclona rubens (Pallas)

Spongia rubens Pallas, 1766, p. 389; Duchassaing and Michelotti, 1864, p. 41, pi. x, fig. 1 ;
Pachychalina rubens, Schmidt, 1870, p. 37 ; Chalina rubens, Carter, 1882, p. 276; Chalinopsis
rubens, Lendenfeld, 1887, p. 744 ; Ceraochalina rubiginosa Lendenfeld 1887, p. 779 ; Pachy-
chalina rubens, Wilson, 1902, p. 392 ; Cladochalina rubens, Burton, 1927, p. 511 ; Haliclona
rubens, de Laubenfels, 1932, p. 59 ; 1936, p. 42, pi. vii, fig. 2, pi. viii, fig. 1 ; de Laubenfels,
1949, P- 9-

Occurrence. St. 31, Gorda Cay, Mosquito Bank, West Indies (15 54' N.,
82 13' W.), 16th November, 1937, 34 m. One specimen.

Distribution. Dry Tortugas ; Bahamas ; Long Key Island ; Guadeloupe ;
Viecques ; St. Domingo ; St. Thomas ; Cuba ; Porto Rico ; 1-17 m.

Haliclona erina de Laubenfels
Haliclona erina de Laubenfels, 1936, p. 457.

Occurrence. St. 10, NW. Arm, Horn Bay, Newfoundland (52 13' N., 55
47' W.), 23rd September, 1937, 15 m. Two specimens.

Distribution. Panama (Caribbean).

Remarks. These specimens, although agreeing with the holotype in structure
and appearance, were obtained well away from the type locality. It is therefore
worth recalling de Laubenfels' comment, in describing the species : "It must be
admitted that this may be a remarkable modification of some previously described
Haliclona, due perhaps to unusual environment conditions, but it is impossible to
say which Haliclona has been so modified." This expresses the situation for so many
specimens of Haliclona from all parts of the world.

224

THE "ROSAURA" EXPEDITION

Haliclona spiculosa (Dendy)

Siphonochalina spiculosa Dendy, 1887, p. 505 ; Dendy, 1890, p. 354, pi. lviii, fig. 2, pi. lxii,
fig- 3 ', Wilson, 1902, p. 394.

Occurrence. St. 24, Turneffe Island, West Indies (17 15' N., 87 49' W.),
1st November, 1937, 2 m. One specimen.

Distribution. Turk's Island, Bahamas ; St. Thomas ; 2-42 m.

Remarks. The species is doubtfully referred to the genus Haliclona, principally
on the absence of a special tangential skeleton at the surface.

Haliclona calcinea sp. n.
(PI. 9, fig. 1 ; Text-fig. 2)

Holotype. B.M. 1938. 6. 30. 31.

Occurrence. St. 34, off St. George, Grenada, West Indies (12 25' N., 6i°
49' W.), 27th October, 1937, 720-800 m. One specimen.

Description. Sponge massive, rounded ; surface even, minutely hispid ; oscules
small, scattered ; texture firm, friable ; colour, in spirit, dark brown ; skeleton
subhalichondroid, mainly unispicular ; megascleres oxea, 0-5 by 0-016 mm.,

Text-fig. 2. Haliclona calcinea sp. n. a, oxeote, x 200 ; b, toxa and sigma, x 500.

THE "ROSAURA" EXPEDITION 225

microscleres centrangulated sigmata, 0-024 mm. chord, and toxa, 0-04 to 0-08 mm.
chord.

Remarks. Although the colour is given as dark brown, this applies strictly to
the ectosome, whether of the outer surface or of the linings of the large canals.
The choanosome is a dull brownish-yellow. The surface, moreover, is largely coated
with a pale greyish yellow, in places up to 3 mm. thick. This apparent incrustation
has the same spicules as the main body and must be presumed to represent new
growth.

The species differs, in its external form, as well as in the detailed measurements of
the microscleres, from other species of Haliclona having such long oxea combined
with the presence of microscleres.

Haliclona tenerrima sp. n.

(PL 9, fig. 2 ; Text-fig. 3)

Holotype. B.M. 1938.6.30.42.

Occurrence. St. 24, Turneffe Island, British Honduras (17 16' N., 87 50' W.),
1st November, 1937, 3J m. One specimen.

Description. Sponge a fragment only ; surface uneven, minutely hispid ;
oscules not present ; texture soft, compressible, elastic ; colour, in spirit, pale

Text-fig. 3. Haliclona tenerrima sp. n. a, oxeote, x 200 ; b, toxon, x 500.

yellow ; skeleton irregularly sub-isodictyal (almost halichondroid) with triangular
mesh, and with occasional fibres of 2 to 3 spicules width running to surface ;
megascleres oxea, 0*28 by 0*007 mm. ; microscleres toxa, 0*03 to o*i mm. chord.

Callyspongia tenerrima Duchassaing & Michelotti

Calyspongia tenerrima Duchassaing & Michelotti, 1864, p. 57, pi. x, figs. 3-4.

Occurrence. St. 31, Gorda Cay, Mosquito Bank, West Indies (15 54' N.,
82 13' W.), 16th November, 1937, 34 m. One specimen.
Distribution. West Indies (St. Thomas, Viecques).

226 THE "ROSAURA" EXPEDITION

Diagnosis. Sponge branching, stipitate, slender, dichotomous ; surface even ;
oscules small, in linear series along branches ; texture soft, compressible ; colour,
in life and in spirit, yellow tinted with rose ; main skeleton an irregular network of
mainly quadrangular mesh, with recognizable primary, secondary and tertiary
fibres ; special ectosomal skeleton, of similar structure but with meshes mainly
triangular ; fibres 0-03 to 0-08 mm. diameter, cored by 1-4 spicules ; megascleres
strongyla, o-i by 0-002 to 0-003 mm. ; microscleres absent.

Remarks. The holotype has been lost, as was ascertained some years ago by de
Laubenfels during his visit to Turin. The original description, inadequate though
it is, leaves, no doubt of the identity of the present specimen. The holotype was
described as " phytoide, dichotome, grele ... les rameaux . . . cylindriques
. . . jaune un peu rose " in life with small oscules in series along the branches.
Every one of these features is found in the present specimen. Above all, the peculiar
appearance of the surface, shown in fig. 4 (Duchassaing & Michelotti, I.e.) puts the
identification beyond reasonable doubt.

Family DESMACIDONIDAE
Section Mycaleae
Genus Desmacella Schmidt
Desmacella annexa (Schmidt)
Desmacella annexa Schmidt, 1870, p. 53 ; Tylodesma annexa, Burton, 1930, p. 525.

Occurrence. St. 34, off St. George, Grenada, West Indies (12 25' N., 6i° 49' W.),
27th November, 1937, 720-800. Two specimens.

Distribution. Iceland ; Norway ; south-west Ireland ; Atlantic coast of
France ; Mediterranean ; West Indies ; Florida ; Indian Ocean ; 85-1331 m.,
on rock, mud and ooze.

Desmacella inornata (Bowerbank)
Halichondria inornata Bowerbank, 1866, p. 271 ; Tylodesma inornata, Burton, 1930, p. 527.

Occurrence. St. 34, off St. George, Grenada, West Indies (12 25' N., 6i°
44' W.), 27th November, 1937, 420-800 m. One specimen.

Distribution. British Isles ; France ; Spain ; Mediterranean ; Azores ; 200-
1025 m., on mud, sand, gravel, broken shells, rocks.

Section Myxilleae
Genus Myxilla Schmidt
Myxilla distort a sp. n.

(PL 9, fig. 3, Text-fig. 4)

Holotype. 1938.6.30.30.

Occurrence. St. 34, off St. George, Grenada, West Indies (12° 25' N., 61 °
49' W.), 27th November, 1937, 720-800 m. One specimen.

THE "ROSAURA" EXPEDITION

227

Description. Sponge erect, lobate ; surface uneven ; harsh to touch ; oscules
at apices of lobes ; texture firm, incompressible, friable ; colour, in spirit, pale
yellow ; main skeleton an irregular reticulation of bluntly-pointed curved styli,
i-o by 0-024 mm. ; ectosomal skeleton a tangential reticulation of irregular mesh
formed of tornota with both ends tylote and measuring 072 by 0-016 mm. ; micro-

Text-fig. 4. Myxilla distorta sp. n. a, style, x 300 ; b, tornote, x 300 ; c, isochela
with teeth meeting in middle line (front view), X 600 ; d, normal large isochela, X 600 ;
e, small isochela, x 600.

zool. 2, 6. 14

228 THE "ROSAURA" EXPEDITION

scleres isochelae spatuliferae, larger frequently distorted by unusual length of teeth,
ranging from 0-024 to 0-08 mm. chord.

Remarks. The species is peculiar in the presence of the distorted chelae, but is
also separable from other known species in the remaining details of its spiculation.

Genus Phorbas Duchassaing & Michelotti

Phorbas amaranthus Duchassaing & Michelotti

Phorbas amaranthus Duchassaing & Michelotti, 1864, p. 92, pi. xxi, fig. 1 ; Cribrella hospitalis
Schmidt, 1870, p. 56, pi. iv, fig. 12 ; ? Anchinoe fictitius, Topsent, 1928, p. 284 ; Phorbas
amaranthus, de Laubenfels, 1936, p. 63.

Occurrence. St. 31, off Gorda Bay, Mosquito Bank, West Indies (15 54' N.,
82 13' W.), 16th November, 1937, 34 m. One specimen.

Distribution. Florida ; West Indies ; ? Azores ; ?? Morocco and Spain ; 27 m.

Remarks. Since this species has never been adequately described, a description
based on the holotype is given :

Sponge encrusting or massive and irregularly digitate ; surface villous, parti-
cularly in larger individuals ; oscules not apparent ; pores in sieves ; texture firm ;
colour, alive, maroon, in spirit, almost white ; skeleton of ascending fibres of serially-
arranged tornota, echinated by acanthostyli of two sizes, and with a tangential
ectosomal layer of tornota similar to those of main skeleton ; tornota, with oxeate
to hastate ends, o-2 to 0-36 by 0-004 to 0-005 mm - '> large acanthostyli, 0-2 by
0-006 mm., small acanthostyli, o-i by 0-006 mm., microscleres isochelae arcuatae,
0-026 mm. chord.

In a very brief redescription of the species, de Laubenfels (1936) describes the
microscleres as chelae and rare sigmata. I have found no sigmata in my preparations
from the type.

Topsent (1928) records specimens of Achinoe fictitius, from the Azores, Morocco
and Spain, having a spiculation similar to that of Phorbas amaranthus. In addition,
he includes Cribrella hospitalis Schmidt as a synonym of his supposed Anchinoe
fictitius, but Schmidt's species is identical with Phorbas amaranthus. Whether,
indeed, this West Indian species extends to the Azores is problematic, but it is highly
doubtful that it should also occur off Morocco and Santander. Re-examination of
Topsent's material is needed to settle this.

Genus Inflatella Schmidt

Inflatella viridis (Topsent)

Joyeuxia viridis Topsent, 1890, p. 29 ; Topsent, 1892, p. 94, pi. ii, fig. 8, pi. x, fig. 19 ; Inflatella
viridis, Lundbeck, 1910, p. 20, pi. ii, figs. 11-12, pi. iv, fig. 7 ; Hentschel, 1929, p. 968 ; nee
Topsent, 1904, p. 205.

Occurrence. St. 34, St. George, Grenada, W 7 est Indies (12 05' N., 6i° 49' W.) f
27th November, 1937, 720-800 m. One specimen.
Distribution. Denmark Strait ; Azores ; 136-1768 m.

THE "ROSAURA" EXPEDITION 229

Genus Plocamionida Topsent

Plocamionida topsenti sp. n.

(Text-fig. 5)

HOLOTYPE. I938.6.3.29.

Occurrence. St. 34, off St. George, Grenada, West Indies (12 25' N., 6i°
49' W.), 27th November, 1937, 720-800 m. One specimen.

Description. Sponge thinly encrusting ; surface hispid ; oscules not apparent ;
colour, in spirit, pale greyish-yellow ; skeleton of styli, 1-2 by 0-04 mm., and acan-
thostyli, 0-24 by 0-008 mm., with occasional intermediates ; ectosomal tornota,
with ends strongylote or faintly subtylote, 0-36 by 0-006 mm., acanthostrongyla,
0-18 by 0-014 mm. ; microscleres isochelae arcuatae, 0-06 to 0-068 mm. chord.

Remarks. The species differs from all others in the genus in the much larger
sochelae, as well as in other details of the spiculation.

Section Tedanieae

Genus Tedania Schmidt

Tedania anhelans (Lieberkuhn)

Halichondria anhelans Lieberkuhn, 1859, p. 521, pi. xi, fig. 6 ; Tedania nigrescens, Burton, 1932,
p. 346, fig. 44 ; Xytopsibis asperus de Laubenfels, 1936, p. 61.

Occurrence. St. 24, Turneffe Island (17 15' N., 8y° 49' W.), 1st November,
1937, 2 m. One specimen.

Distribution. Mediterranean ; Cape Verde Islands ; Bermuda ; West Indies ;
West Africa ; Red Sea ; Indian Ocean ; Indonesia ; Cochin-China ; Australia ;
littoral to 120 m., on rock, stones, shells, mud or sand.

Remarks. The species has been fully discussed (Burton, 1932) under Tedaria
nigrescens (Schmidt) . This name is, however, pre-empted by Halichondria anhelens
Liebekuhn which, as shown by the original description and by Schmidt's (1862,
p. 74) finding of tornota in preparations of the type, is identical with the widespread
species known so long under Schmidt's name.

Family AXINEIXIDAE
Genus Axinella Schmidt
Axinella ramosa sp. n.

(PI. 9, fig. 4 ; Text-fig. 6)

HOLOTYPE . 1 938 . 6 . 30 . 37 .

Occurrence. St. 31, off Gorda Bay, Mosquito Bank, West Indies (15 54' N.,
82 13' W.), 16th November 1937. Three specimens.

23a

THE "ROSAURA" EXPEDITION

^aass^-sfcj^

Text-fig. 5. Plocamionida topsenti sp. n. a, style, x 300 ; b, acanthostyle, x 300 ;
c, acanthostrongyle, x 300 ; d, tornote, x 300 ; e, isochela, x 600.

THE "ROSAURA" EXPEDITION 231

Description. Sponge branching ; surface slightly uneven, minutely hispid ;
oscules not apparent ; texture firm ; colour, in spirit, pale yellow ; skeleton sub-
isodictyal, ascending fibres somewhat plumose, connectives mainly unispicular ;
megascleres styli, 0-4 by 0-016 mm. ; microscleres absent.

Text-fig. 6. Axinella ramosa, sp. n. a, style, x 100.

Remarks. There are three specimens, the largest, the holotype, is 65 mm.
high with its branches 4 mm. in diameter at most. The other two are considerably
smaller.

Genus Hymeniacidon Bowerbank
Hymeniacidon assimilis (Levinsen)

Halichondria assimilis Levinsen, 1886, p. 352, pi. xxx, fig. 5 ; Hymeniacidon assimilis, Burton,
1935. P- 74-

Occurrence. St. 5, Julianshaab, Greenland (6o° 43' N., 46 02' W.), 4th
September, 1937, 80 m. One specimen.

Distribution. Kara Sea ; Sea of Japan ; Sea of Okhotsk ; littoral to 124 m.,
on seaweeds.

Hymeniacidon glabrata sp. n.

(PI. 9, fig. 5 ; Text-fig. 7)

Holotype. 1938.6.30.35.

Occurrence. St. 22, Belize Harbour, West Indies (17 28' N., 88° 11' W.),
29th October, 1937, 6 m. One specimen.

Description. Sponge encrusting, agglutinating ; surface even, minutely hispid ;
oscules at ends of cloacae running parallel to and just under surface ; texture
soft, compressible ; colour, in spirit, pale greyish-yellow ; main skeleton hali-
chondroid, with tendency to multispicular ascending fibres ; ectosomal skeleton
a loose network of more or less quadrangular mesh ; megascleres styli, thickest at
centre (? incompletely differentiated into two sizes), ranging from 0-24 to o-8 by
0-004 to 0-024 mm -

Remarks. The holotype is a thin and flattened, somewhat irregular plate, with
a number of small bivalve shells incorporated in its lower surface. Being no more

232 THE "ROSAURA" EXPEDITION

than a fragment, the interpretation of its mode of growth is difficult. In general
appearance it recalls H. tubulosa (Ridley & Dendy) from the mouth of the River
Plate, which has, presumably, a more erect growth ; and also larger spicules.

Text-fig. 7. Axineila glabrata sp. n. Large and small styli, x 100.

Genus Halichondria Fleming
Halichondria panicea (Pallas)
Spongia panicea Pallas, 1766, p. 388 ; Halichondria panicea, Arndt, 1928, p. 52.

Occurrence. St. 1, Angmagssalik Harbour, Greenland (65 35' N., 37 20' W.),
27th August, 1937, 25-50 m. One specimen.

Distribution. Arctic ; Atlantic coasts of Europe (including Baltic Sea) ;
Mediterranean ; Black Sea ; Atlantic coast of Africa to Saldanha Bay ; St. Paul ;
Patagonia ; Chile ; Atlantic coasts of North America, south to Newfoundland ;
Pacific coast of Asia, south to Japan ; Indian Ocean (Ceylon) ; Kerguelen ; New
Zealand ; littoral to 183 m.

Halichondria tenera (Marenzeller)

Isodictya tenera Marenzeller, 1877, p. 364, pi. i, fig. 2; Burton, 1930, p. 516; Halichondria
tenera, Burton, 1935, p. 76.

Occurrence. St. 1, Angmagssalik Harbour, Greenland (65 35' N., 37 20' W.),
27th August, 1937, 25-50 m. One specimen.

THE "ROSAURA" EXPEDITION 233

Distribution. Franz Josef Land ; Sea of Japan ; Sea of Okhotsk ; 2-178 m.,
mud.

Remarks. This sponge, which is massive, or may be massively branched, with
uneven surface, small scattered oscules and a soft friable texture, has for main
skeleton a loose isodictyal reticulation. The tangential ectosomal skeleton is ill-
developed ; the sole spicules are oxea 0-3 to 0-4 by 0-003 to o-oi mm. The species
is not, therefore, a typical Halichondria. It may even have affinities with the species
of Adocia, except that the oxea are more like those typical for Halichondria.

Halichondria osculum Lundbeck

Halichondria osculum Lundbeck, 1902, p. 23, pi. iii, figs. 3-7, pi. ix, figs. 7-9 ; Hentschel, 1929,
p. 991.

Occureence. St. io, NW. Arm, Horn Bay, Newfoundland (52 13' N.,
55 47' W.), 23rd September, 1937, 15 m. One specimen.

Distribution. Greenland, 18-718 m.

Remarks. The specimen forms a small irregular mass, 10 mm. across, growing
on a Nullipore. There is an encrusting body with processes arising from it and
it is these which determine its identification with Lundbeck's species. Although
much smaller than Lundbeck's original specimens, each process bears a strong
resemblance to them. On the other hand, the spicules are somewhat smaller, none
exceeding o-6 mm., or slightly more, by 0-012 mm., against a maximum of 07
by 0-17 mm., in the types.

On first examination I had taken this to be the type of a new species, but on
further consideration accept it as an atypical specimen of Halichondria osculum.

Halichondria bowerbanki Burton

Spongia coalita Lamouroux, 1816, p. 80; Halichondria coalita, Topsent, 191 1, pp. i-xv ; H.
bowerbanki Burton, 1930, p. 489.

Occurrence. St. 10, NW. Arm, Horn Bay, Newfoundland (17 53' N., 8y°
44' W.), 23rd November, 1937, 15 m. One specimen.

Remarks. Halichondria bowerbanki and H. panicea resemble each other closely
in spiculation and in external appearance. In addition, both are extremely variable
in form. The consequence is that, although Topsent (191 1) has shown marked
differences in their larvae, it is rarely possible to be certain of the identification of a
given specimen. The present sample is of a loose-textured, massive sponge, about
100 mm. across and some 50 mm. high. Its surface is covered with low digitate
processes, 3 to 4 mm. high. The identification given here is based on the translucent
appearance and the fact that specimens having the cavernous quality of the inner
tissues as well as the digitate processes of the surface have been collected by me off
the coast of Devon. Even so, it is given with hesitation.

234

THE "ROSAURA" EXPEDITION

Halichondria cornuloides sp. n.

(PL 9, fig. 6 ; Text-fig. 8)

Holotype. B.M. 1938. 6. 30. 41.

Occurrence. St. 1, Angmagssalik Harbour, Greenland (65 35', 37 20' W.),
27th August, 1937, 25-50 m. One specimen.

Description. Sponge irregularly massive, lobular ; surface smooth, even,
translucent ; oscules (and pores ?) in circular cribriform areas at summits of

Text- fig. 8. Halichondria cornuloides sp. n. Oxea, x 200.

low secondary lobes or apical on main lobes ; texture firm, compressible owing to
cavernous structure of inner tissues ; colour, in spirit, yellowish ; megascleres oxea,
o-2 to 0-5 by 0-008 to 0-012 mm. ; microscleres absent.

Remarks. The species resembles in external appearance the various forms in-
cluded by Dendy in his family Coelosphaeridae, and especially those of the genus
Cornulum. So strong is this resemblance that I found it difficult to be convinced
that it is a true Halichondria. Nevertheless, the oxea have the typical form and
variable length, and the structure of both main and tangential ectosomal skeletons
leave little doubt that it is a Halichondria.

THE "ROSAURA" EXPEDITION

235

Family RASPAILIIDAE

Genus Higginsia Higgin

Higginsia strigilata (Lamarck)

Spongia strigilata Lamarck, 1813, p. 450 ; Lamarck, 18 16, p. 377 ; Higginsia coralloides Higgin,
1877, p. 291, pi. xiv, figs. 1-5.

Occurrence. St. 31, off Gorda Bay, Mosquito Bank, West Indies (15 54' N.,
82 13' W.), 16th November, 1937, 34 m. One specimen.

Distribution. Ireland ; west coast of Africa ; West Indies ; Natal ; Amboina ;
Australia ; littoral (?) to 20 m., on rock or coral rock.

Remarks. Topsent (1933, p. 112) has shown that Higgin's Higginsia coralloides
is the same as Lamarck's Spongia strigilata.

Genus Dragmatyle Topsent

Dragmatyle topsenti sp. n.

(PL 9, fig. 7 ; Text-fig. 9)

Occurrence. St. 26, north of Turneffe Islands, West Indies (17 53' N., 84
44' W.), 7th November, 1937, 900 m. Several fragments (belonging to one
specimen ?)

Text-fig. 9. Dragmatyle topsenti sp. n. Section at right angles to surface, showing
axial core of tylostyli ; surface brushes of styloids, with occasional tylostyli set at
right angles to surface ; and trichodragmata.

Description. Sponge comprising slender branches ; surface uneven, hirsute ;
oscules not apparent ; texture soft, compressible ; colour, in spirit, light brown ;
skeleton an axial core of tylostyli, with occasional tylostyli set at right angles and
projecting beyond surface, with ectosomal brushes of oxeote styloids and tricho-
dragmata for microscleres ; tylostyli i-6 by 0-014 mm., styloids 1-2 by 0-008 mm.
trichodragmata, 0-06 mm. long.

236 THE "ROSAURA" EXPEDITION

Order Keratosa

Genus Hircinia Nardo

Hircinia variabilis Schmidt

Hircinia variabilis Schmidt, 1862, p. 34, pi. iii, fig. 17 ; Lendenfeld, 1889, p. 557, pi. xxxvi,
figs. 11-14.

Occurrence. St. 31, off Gorda Cay, Mosquito Bank, West Indies (15 54' N.,
37 20' W.), 16th November, 1937, 34 m. One specimen.

Distribution. Mediterranean ; West Indies and Florida ; Indian Ocean ;
Australia ; " Pacific Oceanic Islands " ; 6-75 m.

Remarks. If it be possible to recognize varieties in this extremely variable
species, then the present specimen should be assigned to var. dendroides.

LIST OF LITERATURE

Arndt, W. 1928. Porifera, Schwamme, Spongien [in] Tierwelt Dtsch. 4 : 89, no figs.

1935- Porifera [in] Grimpe, Tierwelt N.u. Ostsee, Teil 3a, Lfg. 27 : 1-140, 239 figs.

Arnesen, E. 1901. Spongien fra den norske Kyst. i, Calcarea. Bergens Mus. Aarbo. 5 : 1-

46, pi. i.
— - — 1901. Calcarea Meeresfauna, pp. 65-72.
Bocage, J. V. B. du. 1871. Sur l'existence de la Holtenia carpenteri Wyville-Thompson,

dans les cotes du Portugal. /. Sci. Math. Phys. nat. Lisboa, 3 : 69-70.
Bowerbank, J. S. 1866. A Monograph of the British Spongiadae, 2. London (Ray Society).
Breitfuss, L. 1896. Kalkschwamme der Bremer Expedition nach Ost-Spitzbergen. Zool.

Anz. 19 : 426-432.

1898. Note sur la faune des Calcaires de 1' Ocean Arctique. A nnu. Mus. Zool. Acad.

St. Petersb. 8 : 12-38.

1898. Die Kalkschwamme von Spitzbergen. Zool. Jb. Syst. 11 : 103-120.

191 1. Zur Kenntnis der Spongio-Fauna des Kola-Fjords. Trav. Soc. Nat. St. Petersb.

42 : 223-226.

1932. Die Kalkschwammfauna des arktischen Gebietes. Fauna Arct. Jena, 6 : 235-

252.

Burton, M. 1927. A revision of the genera and species contained in Lendenfeld's " Die
Chalineen des australischen Gebietes." Ann. Mag. nat. Hist. (9) 19 : 289-296, 502-512.

1928. Hexactinellida. Dan. Ingolf-Exped. 6 (4) : 1-18, 9 pis.

1930. Norwegian sponges from the Norman Collection. Proc. zool. Soc. Lond. : 487-

546, 2 pis., 8 figs.

1932. Sponges. " Discovery " Rep. 6 : 327-392, 20 pis., 56 figs.

1934- Sponges. Further zool. Res. Swed. Antarct. Exp. 3 (2) : 1-58. 8 pis., 16 figs.

1935- Some sponges from the Okhotsk Sea and the Sea of Japan. Explor. Mers russes,

22 : 61-79, 6 figs.

Carter, H. J. 1882. Some sponges from the West Indies and Acapulco in the Liverpool
Free Museum. Ann. Mag. nat. Hist. (5) 9 : 266-301, 346-368, 2 pis.

De Laubenfels, M. W. 1932. New sponges from the Puerto Rico deep. Smithson. misc.
Coll. 91 (17) : 1-28.

1936. A discussion of the sponge fauna of the Dry Tortugas. Pap. Tortugas Lab. 30 :

1-225, 22 pis-

THE "ROSAURA" EXPEDITION 237

1949. Sponges of Woods Hole and adjacent waters. Bull. Mus. comp. Zool. Harv.

103 : 1-55, 3 pis.

Dendy, A. 1887. West Indian Chalininae. Proc. zool. Soc. Lond. : 503-507.
— — 1890. Observations on the West Indian sponges. Trans, zool. Soc. Lond. 12 : 349-
368.

& Row, R. W. H. 1 91 1. The classification and phylogeny of the calcareous sponges.

Proc. zool. Soc. Lond. : 704-813.

Derjugin, K. M. 1915. La faune du golfe de Kola. Mem. Acad. Sci. St. Petersb. (8) 35 (1) :

1-929.
Duchassaing, P., & Michelotti, G. 1864. Spongiaires de la mer Caraibe. Nat. Verh.

Holland Maats. Haarlem, 21 : 119 pp., 25 pis.
Ellis, J., & Solander, D. 1786. Natural History of Many Curious and Uncommon Zoophytes.

pp. xii, + 208, 63 pis. London.
Fristedt, K. 1885. Bidrag till Kannedomen om de vid Sveriges vestra Kust lefvande

Spongiae. K. svenska Vetensk-Akad. Handl. (6) 21 : 1-56, pis. i-iv.
Gray, J. E. 1858. On Aphrocallistes, a new genus of Spongiadae from Malacca. Proc.

zool. Soc. Lond. : 11 4-1 15.

1870. Notes on anchoring sponges (Pheronema, Holtenia). Ann. Mag. nat. Hist. (4)

6 : 309-312.

Haeckel, E. 1870. Prodromus eines Systems der Kalkschwamme. Jena. Z. Naturw. 5 :
236-254.

1872. Die Kalkschwamme. 2 vols. & Atlas, 60 pis. Berlin

Hentschel, E. 1916. Die Spongien des Eisfjords. K. svenska. Vetensk-Akad. Handl. 54

(3) : 1-18.

1929. Die Kiesel- und Hornschwamme des Nordlichen Meeres. Fauna arct. Jena,

5 : 859-1042, 4 pis.

Higgin, T. 1877. Description of some sponges obtained during a cruise of the steam yacht

" Argo " in the Caribbean and neighbouring seas. Ann. Mag. nat. Hist. (4) 19 : 291-199,

1 pi.
Hozawa, S. 1 918. Report on the calcareous sponges collected during 1906 by the United

States Fisheries Steamer " Albatross " in the Northwestern Pacific. Proc. U.S. nat.

Mus. 54 : 525-556, 2 pis.

1933. Report on the calcareous sponges obtained by the survey of the Continental

shelf bordering on Japan. Sci. Rep. Tdhoku Univ. 8 (1) : 1-20, 1 pi.

Ijima, I. 1927. The Hexactinellida of the Siboga Expedition. Siboga Exped. 40 (6) : 383 pp.,

26 pis.
Kent, W. S. 1870. On the Hexactinellidae. Mon. Microsc. J. 4 : 241-252.
Lackschewitz, P. 1886. Ueber die Kalkschwamme Menorcas. S. B. naturf. Ges. Dorpat,

7:336-34i-

1886. Ueber die Kalkschwamme Menorcas. Zool. Jb. 1 : 279-310, 1 pi.

Lamarck, J. B. P. de M. 1813. Sur les Polypiers empates. Ann. Mus. Paris, 20 : 294-312,

370-386, 432-458.

1815. Suite des Polypiers empates. Mem. Mus. Paris, 1 : 69-80, 162-168, 331-340.

1816. Histoire naturelle des animaux sans vertebres. 2 Paris, 568 pp.

Lamouroux, J. V. F. 1816. Histoires des Polypiers coralligenes flexibles . Caen, pp. lxxxiv +

558, 19 pis.
Lendenfeld, R. von. 1887. Die Chalineen des Australischen Gebietes. Zool. Jb. 2 : 723-

828, 10 pis.

1889. A Monograph of the Horny Sponges. London. 936 pp. 30 pis.

1903. Tetraxonia [in] Tierreich, 19 : 1-168, 44 figs.

— — 1907. Die Tetraxonia. Wiss. Ergebn. " Valdivia," 11 : 57-373, 46 pis.
Levinsen, G. M. R. 1886. Kara-Havets Svampe (Porifera), pp. 84, 3 pis.

Luberkuhn, N. 1859. Neue Beitrage zur Anatomie der Spongien. Arch. Anat. Physiol,
Lpz. 30 : 293-420, pis. ix-xi.

238 THE "ROSAURA" EXPEDITION

Lundbeck, W. 1902. Homorrhaphidae and Heterorrhaphidae. Dan. Ingolf-Exped. 6 (1) :
108 pp., 19 pis., 1 map.

1906. Porifera (Part III) Desmacidonidae (Pars). Ibid. 6 (3) : 124 pp., 11 pis.

Marenzeller, E. von. 1877. Die Coelenteraten, Echinodermen und Wurmer der K. K.

Oesterreichisch-Ungarischen Nordpol-Expedition. Denkschr. Akad. Wiss. Wien. 35 :

357-398, pis. i-iv.
Marshall, W. 1876. Ideen iiber die Verwandschaftsverhaltnisse der Hexactinelliden. Z.

wiss. Zool. 27 : 1 13-156.
Michlucho-Maclay, N. de. 1868. Beitrage zur Kenntniss der Spongien. I. Ueber Guancha

blanca, einen neuen Kalkschwamm. Jena Z. Naturw. 4 : 221-240.
Muller, O. F. 1776. Zoologiae Danicae, etc. Havniae. xxxii + 282 pp.
Pallas, P. S. 1766. Elenchus zoophytorum, etc. Hagae-Comitis. pp. 28, 451.
Row, R. W. H. 1909. Reports on the marine biology of the Sudanese Red Sea. XIII.

Report on the sponges collected by Mr. Cyril Crossland in 1904-5. Part I. Calcarea.

/. Linn. Soc. Zool. 31 : 182-214, 2 pis., 8 text-figs.
Sars, G. O. 1869. On the deep sea fauna of the Norwegian coast. Vidensk. Selsk. For-

handl. : 250.

1872. On Some Remarkable Forms of Animal Life from the Great Deeps off the Norwegian

Coast. Christiania. 82 pp., 6 pis.

Schmidt, E. O. 1862. Die Spongien des adriatischen Meeres. Leipzig. 44 pp., 5 pis.

1870. Grundziige einer Spongien- Fauna des atlantischen Gebietes. Leipzig. 88 pp.,

6 pis.

1880. Die Spongien des Meerbusens von Mexico und des Caraibischen Meeres. Heft 2.

Jena. 33-90, pis. v-x.

Schuffner, O. 1877. Beschreibung einiger neuer Kalkschwamme. Jena. Z. Naturw.

11 : 403-433, pis. xxiv-xxvi.
Schulze, F. E. 1886. Ueber den Bau und das System der Hexactinelliden. Abh. preuss.

Akad. Wiss. : 1-97.

1887. Report on the Hexactinellida collected by H.M.S. " Challenger." Rep. Sci.

Res. " Challenger," Zool. 21 (53) : 514 pp., 104 pis. 1 map.

1893. Revision des Systemes der Hyalonematiden. S.B. Preuss. Akad. Wiss. 30 :

54I-589.

1904. Hexactinellida. Wiss. Ergebn. " Valdivia," 4 : viii + 226 pp., 52 pis.

Sollas, W. J. 1886. Preliminary account of the Tetractinellia sponges dredged by H.M.S.

" Challenger." Sci. Proc. R. Dublin Soc. 5 : 177-199.

1888. Report on the Tetractinellida. Rep. Sci. Res. Voy. " Challenger," Zool. 25 :

clxvi, + 458, 54 pis.

Thacker, A. G. 1908. On collections of the Cape Verde Islands fauna made by Cyril Cross-
land. Proc. zool. Soc. Lond. : 757-782, 1 pi., 12 text-figs.

Thiele, J. 1898. Studien iiber pazifischen Spongien. Zoologica, Stuttgart, 24 : 1-72, 8 pis.

Thomson, C. W. 1869. On Holtenia, a genus of vitreous sponges. Ann. Mag. nat. Hist.
(4) 4 : 284-287.

Topsent, E. 1890. Notice preliminaire sur les Spongiaires receuillis durant les campagnes de
l'Hirondelle. Bull. Soc. zool. Fr. 15 : 26-32, 65-71.

1892. Contributions a l'etude des Spongiaires de l'Atlantique Nord. Result. Camp.

sci. Monaco, 2 : 165 pp., 11 pis.

1904. Spongiaires des Acores. Result. Camp. sci. Monaco, 25 : 280 pp., 18 pis.

191 1. Sur les afhnites des Halichondria et la classification des Halichondrines d'apres

leurs formes larvaires. Arch. Zool. exp. gen. (5) 7 : i-xv, 3 figs.

1913. Spongiaires provenant des campagnes scientifiques de la " Princess Alice " dans

les Mers du Nord. Result. Camp. sci. Monaco, 45 : 1-67, 5 pis.

1922. Sur Stelletta crassispicula Sollas et son synonyme Stelletta crassiclada Lendenfeld.

Bull. Inst, oceanogr. Monaco, No. 416 : pp. 1-3.

THE "ROSAURA" EXPEDITION

239

1928. Spongiaires de l'Atlantique et de la Mediterranee. Result. Camp. sci. Monaco.

74 : 376 pp., 11 pis.

1934- sponges observees dans les parages de Monaco. Bull. Inst, oceanogr. Monaco,

650 : 1-42, 3 figs.

Wilson, H. V. 1902. The Sponges collected in Porto Rico. Bull. U.S. Fish. Comm. 2 :
375-411, 30 figs.

1904. Reports on an exploration off the West Coasts of Mexico, Central and South

America, and off the Galapagos Islands. Mem. Haw. Mus. comp. Zool. 30 (1) : 1-164,
26 pis.

PRESENTED
5 OCT 1954

PLATE 9

Fig. i. Haliclona calcinea sp. n. x i/i.

Fig. 2. Haliclona tenerrima sp. n. X 3/2.

Fig. 3. Myxilla distorta sp. n. x 1/1.

Fig. 4. Axinella ramosa sp. n. x 1/1.

Fig. 5. Hymeniacidon glabrata sp. n. x 1/1.

Fig. 6. Halichondria cornuloides sp. n. x 1/1.

Fig. 7. Dragmatyle topsenti sp. n. X 3/2.

Bull. B.M. {N.H.) Zool. 2, 6.

PLATE 9.

PRESENTED

5 GOT 1954

PRINTED IN GREAT BRITAIN BY
ADLARD AND SON, LIMITED,
BARTHOLOMEW PRESS, DORKING

NOV 1954

ON THE POLYZOAN GENUS

CREPIDACANTHA

LEVINSEN

D. A. BROWN

BULLETIN OF
THE BRITISH MUSEUM (NATURAL HISTORY)
ZOOLOGY Vol. 2 No. 7

LONDON: 1954

ON THE POLYZOAN GENUS CREPIDACANTHA

LEV1NSEN

DAVID ALEXANDER BROWN v . /

(Geology Department, University of Otago, Dunedin, New Zealand)

Pp. 241-263 ; 4 Text-figures

BULLETIN OF

THE BRITISH MUSEUM (NATURAL HISTORY)

ZOOLOGY Vol. 2 No. 7

LONDON : 1954

THE BULLETIN OF THE BRITISH MUSEUM
(NATURAL HISTORY), instituted in 1949, is
issued in five series, corresponding to the Departments
of the Museum, and an Historical Series.

Parts appear at irregular intervals as they become
ready. Volumes will contain about three or four
hundred pages, and will not necessarily be completed
within one calendar year.

This paper is Vol. 2, No. 7 of the Zoological Series.

PRINTED BY ORDER OF THE TRUSTEES OF
THE BRITISH MUSEUM

Issued October, 1954 Price Six Shillings

ON THE POLYZOAN GENUS CREPIDACANTHA

LEVINSEN

By DAVID ALEXANDER BROWN

SYNOPSIS

The family Crepidacanthidae is considered to include only the genus Crepidacantha Levinsen.
The known species of that genus, including two new ones, are described, and their geographic
distribution discussed.

INTRODUCTION

During a study of the New Zealand Tertiary Cheilostomatous Polyzoa (Brown,
1952), in which the family Crepidacanthidae is moderately well represented, I have
made many comparisons with material from other parts of the world. The results
and conclusions concerning that family, which are beyond the scope of the Catalogue,
are now put forward. Canu & Bassler (1929 : 409-412 ; 1930 : 32, 33) and Marcus
(1938 : 233) have already dealt briefly with the species of Crepidacantha, but as
their results, so far as the Indo-Pacific fauna was concerned, were probably not
based on actual material, a good deal of confusion has occurred. This, I think, may
now be rectified after an examination of the material in the British Museum (Natural
History).

ACKNOWLEDGMENTS

My thanks are due to the Keepers of Zoology and Geology in the British Museum
(Natural History) for facilities provided for my research, to the Director of Universi-
tetets Zoologiske Museum, Copenhagen, for the loan of specimens from the Levinsen
Collection, and to Dr. Anna B. Hastings and Dr. H. Dighton Thomas of the British
Museum (Natural History) for their helpful criticisms of this paper.

FAMILY CREPIDACANTHIDAE

Levinsen's original definition of this family (1909 : 266) is quite adequate, though
it is clear that the words " distal half " should be substituted for " proximal half "
in regard to the position of the " 9-12 very long marginal spines," and that " hetero-
zooecia " is a better term than " vibracula " in view of the large chambers occurring
in C. zelanica Canu & Bassler which often appear to be avicularian in nature.

Genus Crepidacantha Levinsen

Crepidacantha Levinsen, 1909, Morph. Syst. Stud. Cheil. Bry. : 266 ; Brown, 1952 : 359
(cum syn.).

244 ON THE POLYZOAN GENUS C RE P I D AC A N T H A LEVINSEN

Type species (by monotypy). 1 Crepidacantha poissonii (Audouin) var. crinispina
Levinsen, 1909 : 266, 267, text-figs. 1, 2, 3, 5, 6. Recent : Thailand.

The orifice of this genus has a wide poster, and the recumbent 2 ovicell has an
ectooecium with an uncalcified frontal area of variable size in which the entooecium,
which usually has pores, is visible. Marginal spines occur along the distal-lateral
edges of the zooecia alternating with the pore-chambers which are often marked by
slit-like areolae. Paired heterozooecia 3 , usually vibracular in nature, with long
setiform flagella, are present, either placed as tubular chambers on each side of the
orifice or as rounded chambers on the frontal wall proximally to the orifice.

Bassler (1935 : 81) records the genus as ranging from Cretaceous to Recent, but
I have not been able to find any records of species earlier than the Waiauan [Middle
Miocene] stage of New Zealand. However, as it is even then axlearly distinct
genus, it will probably be found in beds of earlier age. From the evidence available,
it appears to be mainly a tropical and sub-tropical encrusting genus ranging from
0-250 fathoms, though specimens have been found in New Zealand waters as far
south as Kaka Point (169 50' E., 46 25' S.).

The orifice of Crepidacantha is evidently similar in construction to that of the family
Hippoporinidae though rather peculiar in the breadth of the shallow poster. The
latter may vary considerably in shape, with a convex proximal lip as in C. solea
Canu & Bassler (Fig. ih) or a very concave one as in C. setigera Smitt sp. (Fig. ie).
However, a separate family appears to be justified in view of the nature of the
orifice and of the constant occurrence of paired, symmetrically placed heterozooecia.

Canu & Bassler (1929 : 408) apparently considered the presence of a recumbent
ovicell as a family character for they grouped the genera Mastigophora Hincks
and Schizobathy sella Canu & Bassler with Crepidacantha on this account (see also
Canu & Bassler, 1927 : 21, where P achy kr asp edon Koschinsky, Nimba Jullien and
Nimbella Jullien are also included in the Crepidacanthidae ; and Canu & Lecointre,
1930 : no). An examination of these genera, however, shows clearly (from the
nature of the orifice, which is, in my opinion, of far greater importance from the
point of view of classification than the presence of a recumbent ovicell) that they
can be grouped neither with Crepidacantha nor together. Mastigophora Hincks
(= Escharina Milne-Edwards) and Schizobathy sella Canu & Bassler belong to the
Schizoporellidae. Nimba Jullien and Nimbella Jullien appear to be closely related,
and, though Jullien & Calvet (1903) placed the former in the Schizoporellidae and
the latter in the Sertellidae (Reteporidae), they should probably be placed together

1 It may be argued that Levinsen also included the species C. poissonii itself as a genosyntype. If
this be accepted, then the selection of type must be Flustra poissonii Audouin, 1826, p. 240, pi. 10,
figs. 5 1 , 5 2 , by Edwards (1910 : 21).

2 This term appears to have been used first by Canu & Bassler (191 7 : 66) and later amplified in their
1920 monograph (pp. 54, 55, text-fig. 10, " Independent (recumbent) aneucleithrian hyperstomial
ovicell "). The peculiarity of this type of ovicell may be seen at the edge of a colony. Here, the ovicell
is observed to be attached to the distal wall of the fertile zooecium and is fully formed before any trace
of distal zooecia appears. The latter then grow round the ovicell so that it appears more or less embedded
in them. In the case of the ordinary hyperstomial ovicell, the frontal wall of the distal zooecium is in
process of growth at the same time as the ovicell of the proximal zooecium, the rudiment of the ovicell
making its appearance on the proximal part of the frontal wall of the developing distal zooecium (see
Levinsen, 1909, pi. 19, fig. 4a).

3 The ancestrula on a specimen of C. altirostris Canu & Bassler ( = C. crinispina Levinsen sp.) appeared
to possess a single vibracular chamber (see under C. crinispina).

ON THE POLYZOAN GENUS C RE P I D A C A N T H A LEVINSEN

245

in a separate family. P achy kr asp edon Koschinsky probably belongs also to the
Schizoporellidae as its primary orifice possesses a deep, often sub-circular sinus
flanked by a pair of stout condyles ; there is a thick raised peristome and a coarsely
perforate frontal wall. I therefore regard the Crepidacanthidae as including only
one genus, as did Levinsen.

0.25mm.

B-J

Q25mm.

Fig. 1. Species of Crepidacantha showing orifice, ovicell and position of heterozooecia.
a. Crepidacantha poissonii (Audouin) var. teres Hincks. Oblique view showing
structure of ectooecium. B.M., 11. 10. 1.783. Recent: Madeira, b. The same,
frontal view. c. C. crinispina (Levinsen). Lectotype. Recent: Bangkok, Thai-
land, d. C. crinispina (Levinsen) var. parvipora (Canu & Bassler). B.M., D. 36894.
Waiauan [Middle Miocene] : Weka Pass, New Zealand, e. C. setigera (Smitt). B.M.,
32.3.7.27. Recent: Straits of Florida, f. C. longiseta Canu & Bassler. B.M.,
99. 7. 1. 1 722. Recent: John Adams Bank, Brazil, g. C.grandis Canu & Bassler.
B.M., 31. 12.30. 158. Recent: Philippines, h. C. solea Canu & Bassler. Lectotype,
B.M., 89. 8. 21. 19. Recent: Tizard Reef, China Sea. j. C. zelanica Canu & Bassler.
B.M., D. 36897. Nukumaruan [Middle Pliocene] : Petane, New Zealand.

zool. 11, 7.

*5§

246 ON THE POLYZOAN GENUS CRE P I D A C A NT H A LEVINSEN

Vigneaux's contention (1949 : 19) that the family Crepidacanthidae is superfluous
and that Crepidacantha, along with a number of other genera, should be placed in
the family Phylactellidae 1 appears to be quite without foundation. In so far as
Crepidacantha is concerned, the only point of resemblance to Phylactella Hincks,
1879 [genolectotype (chosen by Canu & Bassler, 1917 : 67) : Lepralia labrosa Busk,
1854 : 65, 82, pi. 92, figs 1, 2, ? 3] is the possession of a recumbent ovicell.

Livingstone (1929 : 94) has stated that the position of the vibracula is of little
importance in classifying the species of Crepidacantha. However, the orientation
of the flagella (i.e., laterally or proximally) does appear to be constant within a
species, and it has been found, with one exception, viz., C. crinispina (Levinsen)
var. parvipora Canu & Bassler var., that where the vibracular chambers are placed
proximally to the orifice, the flagella are directed laterally inwards, whereas those
species with vibracular chambers alongside the orifice have their flagella directed
proximally. It is found also that these differences in the orientation of the flagella
are accompanied by other important distinguishing characters such as the structure
of the ovicell and the shape of the zooecial orifice.

key to the species of Crepidacantha described here

1 . Ovicell with uncalcified area of ectooecium occupying most of the frontal area . 2
Ovicell with uncalcified area of ectooecium much less extensive . . . .10

2. Heterozooecia placed alongside orifice. Flagella directed proximally ... 3
Heterozooecia placed proximally to orifice. Flagella usually directed laterally inwards. 5

3. Ovicell large. Poster very deep, concave .... e. C. setigera (Smitt)
Ovicell not large. Poster shallow, convex ........ 4

4. Proximal lip of orifice square and plate-like. Calcified ectooecium marginal

c. C. crinispina (Levinsen)
Proximal lip of orifice rounded, convex. Calcified ectooecium wide

g. C. grandis Canu & Bassler

5. Heterozooecia large. Ovicell coarsely perforate . .J. C. zelanica Canu & Bassler
Heterozooecia small ............ 6

6. Median frontal umbo much enlarged. Flagella directed proximally

d. C. crinispina (Levinsen) var. parvipora (Canu & Bassler)
Median tubercle small or absent. Flagella directed inwards ..... 7

7. Entooecium with longitudinal radiating ridges ....... 8

Entooecium without radiating ridges, but may have median carina ... 9

8. Proximal lip of orifice convex, rounded . . . a. C. poissonii (Audouin)
Proximal lip of orifice convex, square, plate-like . . h. C. solea Canu & Bassler

9. Ovicell deeply immersed. Calcified ectooecium marginal. Proximal lip of orifice convex

square, plate-like ....... l. C. kirkpatricki, sp. nov

Ovicell globular. Calcified ectooecium covering proximo-lateral portions of ovicell. Proximal

lip of orifice nearly straight . . . b. C. poissonii (Audouin) var. teres Hincks

10. Entooecium exposed in depressed transverse oval area . f. C. longiseta Canu & Bassler

Entooecium exposed in raised longitudinal oval area . k. C. bracebridgei, sp. nov

a. Crepidacantha poissonii (Audouin)

Flustra poissonii Audouin, 1826, Description de l'Egypte : 240 ; Savigny, 1826, pi. 10, figs.

5 1 , 5 2 ; Audouin, 1828 : 68.
Reptescharellina poissonii d'Orbigny, 1852, Paleontologie Francaise. Terr. Cret. 5 : 454.
Crepidacantha poissoni, Canu & Bassler, 1929, Bull. U.S. nat. Mus. 100 (9) : 409, text-fig. i6oa.

1 See also Canu & Bassler, 1923 : v.

ON THE POLYZOAN GENUS CREP ID AC ANT H A LEVINSEN 247

Diagnosis (from Savigny's figure). Crepidacantha with entooecium ornamented
front ally with longitudinal radiating bars. Vibracula placed proximally to orifice.
Proximal lip of orifice convex, rounded.

Remarks. Audouin (1826 : 240 ; 1828 : 68) in his notes on the species figured
in Savigny's plate 10, gives no locality for these, and they may, therefore, be either
from the Red Sea or the Egyptian coast of the Mediterranean 1 . I have not seen
any specimens from either of these localities 2 and I am unable to decide whether
this species is the same as the variety from Madeira (see below). Savigny has,
however, clearly figured radiating longitudinal ridges on the frontal surface of the
entooecium and these do not appear on any of the specimens examined from Madeira.
The proximal lip of the orifice of Audouin's species is also much more convex than
that of the Madeiran material, which I have recognized, therefore, as a distinct
variety, teres Hincks. It is unfortunate that Savigny did not indicate in his figure
the orientation of the vibracular setae, but it is most probable that, as the vibracular
chambers are placed proximally to the orifice, the flagella were directed laterally
inwards.

Distribution 1 . ? Eastern Mediterranean ; ? Red Sea.

b. Crepidacantha poissonii (Audouin) var. teres Hincks

(Fig. IA, B)

Lepralia kirchenpaueri Heller var. teres Hincks, 1880, Ann. Mag. nat. Hist. (5) 6 : 77, pi. 9, figs.

7, ya ; Hincks, 1891 : 88, 89.
Lepralia teres Hincks, 1895, Index Mar. Polyzoa : ii, note.
Lepralia poissonii Audouin, Waters, 1899, J.R. micr. Soc. : 16 ; Norman, 1909 : 307, pi. 41,

figs. 7, 8 ; L. poissoni, Calvet, 1907 : 409.
Crepidacantha poissoni, Canu & Bassler, 1929, Bull. U.S. nat. Mus. 100 (9) : 409, text-figs. i6ob,

c ; C. poissonii, Osburn, 1940 : 451 ; ? Crepidacantha poissonii, Canu & Bassler, 1928a : 136,

pi. 34, fig.3.

Material examined : —

B.M., 11. 10. 1. 783 s (Figs, ia, b). Recent : Madeira. Specimen encrust-
ing small Pecten. Labelled "Lepralia Poissonii Aud. Madeira, 1897. A.M.N."
Norman Collection. '

B.M., 11. 10. 1. 784. Recent: Madeira. Labelled "Lepralia Poissonii
Audouin. Salvages. Senor D. Noronha." Norman Collection.

B.M., 11. 10. 1.785. Recent: Madeira. Labelled "Lepralia Poissonii Aud.

Porto Santo. Senor de Noronha. 1908." Norman Collection.

Diagnosis. Crepidacantha with the calcified ectooecium confined to the lateral

margins of the ovicell, the entooecium with a transverse porous ridge. Vibracular

flagella directed laterally. Proximal lip of orifice straight or slightly convex.

Marginal spines slender, about twelve in number.

1 D'Orbigny (1852 : 454) cites the Red Sea for C. poissonii. See also Postscript.

2 Monsieur E. Buge of the Laboratoire de Paleontologie, Museum National d'Histoire Naturelle,
Paris, states (in Hit.) that the Savigny & Audouin Collection is considered lost.

3 Specimens in the British Museum (Natural History) are referred to by their Register Numbers.

248 ON THE POLYZOAN GENUS C RE P I D A C A N T H A LEVINSEN

Remarks. The chief differences between this variety and Audouin's species
have been noted above. Another important character is the proximo-lateral
extension of the calcified ectooecium which covers the entooecium near the orifice.
The ridge with the row of pores traversing the frontal wall of the ovicell does not
mark the edge of the calcareous ectooecium, as I thought at first (Brown, 1952 : 361),
but is an ornamentation of the entooecium.

Distribution. Recent : Madeira ; Cape Verde Islands, 1 10-180 metres (Calvet) ;
Porto Rico, 6 fathoms (Osburn) ; Bermuda (Osburn).

Fossil : ? Pliocene : Bocas Island, NW. Panama (Canu & Bassler).

c. Crepidacantha crinispina (Levinsen)
(Fig. ic)

Lepralia poissonii Audouin, Waters, 1887, Q. J. geol. Soc. Lond. 43 : 42, 59 (part. — non pi. 8,
fig. 37 = C. zelanica Canu & Bassler) ; Philipps, 1899 : 440, 446 ; ? Lepralia poissonii,
Hincks, 188 1 : 122 ; ? Lepralia poissonii (second form), Hincks, 1885 : 256.

Lepralia setigera Smitt, MacGillivray, 1883, Trans, roy. Soc. Vict. 19 : 133, pi. 1, figs. 2, 3 ;
MacGillivray, 1887 : 212.

? Crepidacantha poissoni Livingstone, 1929, Vidensk. Medd. dansk naturh. Foren. Kbh. 87 : 93.

Crepidacantha poissonii Audouin var. crinispina Levinsen, 1909, Morph. Syst. Stud. Cheil.
Bry. : 266, text-figs. 1, 2, 3, 5, 6 ; C. poissoni var. crinispina, Canu & Bassler, 1923 : 174,
text-figs. 33A-C, e, f.

Crepidacantha setifera Canu & Bassler, 1929, Bull. U.S. nat. Mus. 100 (9) : 409 ; Canu & Bassler,

1930 : 32, 33-
Crepidacantha papulij 'era Canu & Bassler, 1929, Bull. U.S. nat. Mus. 100 (9) : 410, text-figs. i6ok,

i6ib, pi. 57, fig. 8.
Crepidacantha crinispina (Levinsen), Canu & Bassler, 1929, Bull. U.S. nat. Mus. 100 (9) : 409,

text-figs. i6od-f, h, 1 ; Brown, 1952 : 359, text-fig. 283 ; Brown, 1954 : 433.
Crepidacantha altirostris Canu & Bassler, 1929, Bull. U.S. nat. Mus. 100 (9) : 411, pi. 57, fig. 9.

Lectotype (chosen by Brown, 1952 : 359). Universitetets Zoologiske Museum,
Copenhagen, specimen encrusting fragment of large lamellibranch (Fig. ic). Recent :
Koh Kram, Bangkok, Thailand, 30 fathoms. Levinsen Collection.
Other material examined : —

B.M., 97.5.1.832, 835. Recent: Port Phillip Heads, Victoria. Two
encrusting specimens. Labelled "Lepralia setigera." Bracebridge Wilson
Collection.

B.M., 88 . 11 . 14 . 307. Recent : Port Phillip, Victoria. Labelled " Lepralia
poissonii Aud." by Kirkpatrick. Bracebridge Wilson Collection.

B.M., 99.5.1.825. Recent: Wellington, New Zealand. Labelled
" Lepralia Poissonii Audouin" by Miss Jelly. Hincks Collection.

B.M., 31. 12. 30. 157. Recent: Albatross Stn. 5179, 37 fathoms off
Romblon Light, Romblon, Philippines. Paratype of C. altirostris Canu &
Bassler. Presented by the U.S. National Museum.

B.M., D. 36893. Castleclifhan CU3 fUpper Pliocene] : CastleclifT, New
Zealand. Sent by Mr. C. A. Fleming.

Waitotaran [Middle Pliocene] : submarine limestone off Three Kings
Islands, New Zealand.

ON THE POLYZOAN GENUS C RE P I D A C A N T H A LEVINSEN 249

Diagnosis. Crepidacantha with the calcified ectooecium confined to the distal-
lateral margins of the ovicell. Entooecium with a longitudinal median carina.
Vibracula placed alongside the opercular condyles, the flagella directed proximally.
Proximal lip of orifice convex, often plate-like. Frontal wall of zooecium often
bearing a prominent sub-oral tubercle. Marginal spines ten in number.

Remarks. In addition to the comments already made on this species by Brown
(1952 : 360) the following remarks are apposite.

This species is clearly distinct from both C. poissonii (Audouin) and from C.
poissonii var. teres in the different construction of its ovicell and in the position and
orientation of its vibracular setae.

In the lectotype the ovicells have marked e version of the transverse ridge, which
tends to be masked in the Australian and New Zealand specimens by thickening of
the ectooecium.

Although MacGillivray (1883 : 133, pi. 1, fig. 2) does not show the median carina
on the ovicell of his Lepralia setigera Smitt, there can be little doubt that this is the
species described by him, as it differs completely from the other species of Crepi-
dacantha from Port Phillip, Victoria, namely, C. bracebridgei , sp. n., and C. kirkpatricki ,
sp. n. This identification is substantiated by two specimens from Port Phillip
Heads sent by Mr. J. Bracebridge Wilson and labelled by him "Lepralia setigera "
(B.M., 97.5.1.832, 835). Canu & Bassler's name, C. setifera (1929 : 409), proposed
for Lepralia setigera MacGillivray, must therefore lapse in favour of Levinsen's
name, C. crinispina.

The specimen (B.M., 88. n . 14.307) from Port Phillip, Victoria, has a very marked
sub-oral umbo resembling that of C. crinispina (Levinsen) var. parvipora Canu &
Bassler var. from New Zealand, but the plate-like proximal lip of the orifice and the
rather distally placed vibracula distinguish it from that variety.

In the Recent specimen (B.M., 99.5.1.825) from Wellington, New Zealand, the
vibracular chambers are rather larger than those of the lectotype and the sub-oral
tubercle is often absent : it is probably C. crinispina.

Canu & Bassler's reasons for separating C. papulifera (1929 : 409, text-fig. 160K ;
410, 411, text-fig. 161B, pi. 57, fig. 8) from C. crinispina appear to be without
foundation.

C. altirostris Canu & Bassler (1929 : 411, pi. 57, fig. 9) is represented in the British
Museum (Natural History) by a paratype (B.M., 31. 12. 30. 157) and appears to be
conspecific with C. crinispina. The vibracula are sometimes placed near the distal
end of the zooecium as stated by Canu & Bassler, but this feature is also seen in the
lectotype of Levinsen's species. The only difference to be seen in the specimen of
C. altirostris is the rather more marked carination of the entooecium which shows
no e version of the transverse ridge. The ancestrula is preserved in this specimen.
It is small and appears to have a single vibraculum placed proximally and to one
side of the orifice.

Distribution. Recent : Thailand (30 fathoms) ; Philippines (21-105 fathoms) ;
Port Phillip, Victoria ; Loyalty Islands, New Caledonia (35 fathoms) (Philipps) ;
? Bass Strait (Hincks) ; New Zealand (3-10 fathoms) .

Fossil : Castlecliffian [Upper Pliocene] : Castlecliff ; ? Shakespeare Cliff, New

250 ON THE POLYZOAN GENUS C RE P I D A C A N T H A LEVINSEN

Zealand. ? Nukumaruan [Middle Pliocene] : Waipukurau, New Zealand (Waters).
Waitotaran [Middle Pliocene] : submarine limestone off Three Kings Islands, New
Zealand.

d. Crepidacantha crinispina (Levinsen) var. parvipora (Canu & Bassler)

(Fig. id)

Lepralia poissonii Audouin (third form), Hincks, 1885, Ann. Mag. nat. Hist. (5) 15 : 256 ; L.

poissonii, Waters, 1887 : 59 (part. — non pi. 8, fig. 37 = C. zelanica Canu & Bassler).
Crepidacantha parvipora Canu & Bassler, 1930, Proc. U.S. nat. Mus. 76 (13) : 32, 33.
Crepidacantha crinispina (Levinsen) var. parvipora, Brown, 1952, Tert. Cheil. Polyzoa N.
Zealand : 361, text-rig. 284.

Neotype (chosen by Brown, 1952 : 361). B.M., D. 37042. [Pliocene] : [Napier],
New Zealand. Labelled " Lepralia Poissonii var. R. Tert. N. Zealand" by Miss
Jelly. Hincks Collection.
Other material examined : —

B.M., D. 36777. Nukumaruan [Middle Pliocene] : Waipukurau Gorge.
Encrusting specimen. Hincks Collection.

B.M., D. 32531. Nukumaruan [Middle Pliocene] : Waipukurau Gorge.
Encrusting specimen. Labelled "Lepralia Poissonii Aud." by Miss Jelly.
Hincks Collection.

B.M., D. 36894 (Fig. id)-36896. Waiauan [Middle Miocene], Base of
Uppermost Mt. Brown " E " Limestone : Junction of Weka Creek and Weka
Pass Stream, Waipara, S.D. Three encrusting specimens. Sent by Pro-
fessor B. H. Mason.
Diagnosis. C. crinispina with a stout median ridge on the frontal wall ending
distally in a sub-oral umbo. Vibracula placed on either side of the umbo, the
flagella directed proximally.

Remarks. This variety has been fully discussed by Brown (1952 : 361). Unlike
the other known species of Crepidacantha in which the vibracula are placed proxi-
mally to the orifice, this variety has the flagella directed proximally. The presence
of the marked sub-oral umbo would inhibit or prevent the inward turning of the
vibracula.

Distribution. Fossil : Nukumaruan [Middle Pliocene] : [Napier] ; Waipu-
kurau Gorge, New Zealand. Waiauan [Middle Miocene] : Weka Pass, New Zealand.

e. Crepidacantha setigera (Smitt)

(Fig. ie)

Eschar ella setigera Smitt, 1873, K. svensk. Vetensk. — Akad. Handl. 11 (4) : 58, 75, 82, pi. 10, fig. 206.
Crepidacantha setigera, Canu & Bassler, 1928a, Proc. U.S. nat. Mus. 72 (14) : 135, pi. 21, fig. 10 ;
Canu & Bassler, 1929 : 409 ; Osburn, 1940 : 452 ; Osburn, 1952 : 479, pi. 58, fig. 1.

Material examined : —

B.M., 32.3.7.27 (Fig. ie). Recent : Albatross Stn. 2639, Straits of Florida.
Two encrusting specimens. Presented by the U.S. National Museum.
Diagnosis. Crepidacantha with large ovicell not clearly separated from the
zooecium, the entooecium with a curved row of pores, the ectooecium wide.

ON THE POLYZOAN GENUS CREP I D A C A N T H A LEVINSEN 251

Vibracula placed just above the level of the opercular condyles, the flagella directed
proximally. Poster very deep and concave.

Remarks. This species is very clearly distinguished by the shape of its orifice
and its large ovicell. Smitt (1873 : 57, 82) regarded it as a variety of Lepralia
depressa Busk (1854 : 75, pi. 91, figs. 3, 4), to which, however, it does not seem to be
related.

Osburn's record (1952: 479) of this species from the Galapagos Islands is interesting.
Whether the distribution of this and* other Polyzoan species, now living on both
sides of the Isthmus of Panama, indicates a geologically recent connection or is
due to transport of larvae and colonies on, for instance, the bottoms of ships
traversing the canal, will probably never be known. It is likely, however, that the
fresh- water Gatun Lake would be an efficient barrier to such a passage.

Distribution. Recent : Gulf of Mexico (50-60 fathoms) : Galapagos Islands
(Osburn).

f. Crepidacantha longiseta Canu & Bassler
(Fig. if)

Lepralia poissonii Audouin, Kirkpatrick, 1888a, Ann. Mag. nat. Hist. (6) 1 : 78, pi. 8, fig. 1 ;

? L. poissonii, Thornely, 1912 : 150 ; Waters, 1914 : 856.
Crepidacantha longiseta Canu & Bassler, 1928a, Proc. U.S. nat. Mus. 72 (14) : 135, pi. 21, figs.

3, 4 ; Osburn, 1940 : 452.
Crepidacantha levinseni Marcus, 1938, Vidensk. Medd. dansk naturh. Foren. Kbh. 101 : 231

(part.), text-figs. 2 8a, b.

Holotype. U.S. National Museum No. 7826. Recent : Cuba.

Material examined : —

B.M., 88. 1. 25. 31. Recent: Mauritius. Labelled "Lepralia Poissonii
Audouin." V. Robillard Collection.

B.M., 34.10.6.19. Recent: Mauritius.

B.M., 99.7. 1. 1722 (Fig. if). Recent : John Adams Bank, South Atlantic,
off Brazil. Labelled " Lepralia Poissonii Aud." Busk Collection.

Diagnosis. Crepidacantha with porous entooecium visible through a transverse,
sub-circular or oval fenestra in the ectooecium. Vibracular chambers tubular,
placed at the level of the opercular condyles, the flagella very long and directed
proximally. Proximal lip of orifice straight or slightly concave. Marginal spines
ten in number.

Remarks. I have not examined any West Indian material of C. longiseta, but
a specimen (B.M., 99.7.1. 1722, Busk Coll., collected by H.M.S. " Herald" from
the John Adams Bank off the coast of Brazil), undoubtedly represents that species.
This specimen also agrees closely with the lectotype {here chosen) of Marcus's C.
levinseni from St. Helena (viz., the specimen figured by him, 1938, text-fig. 28A on
p. 232), except that the porous area (entooecium) in the ectooecium of the Brazilian
specimen appears to be quite separated from the orifice (Fig. if). As shown later,
the species from Port Phillip, Victoria, with which Marcus correlated his C. levinseni,
is new, namely, C. kirkpatricki.

252 ON THE POLYZOAN GENUS C RE P I D A C A N T H A LEVINSEN

The original material from Mauritius described by Kirkpatrick as Lepralia
Poissonii Audouin (B.M., 88. 1.25. 31), and another specimen from the same
locality (B.M., 34. 10.6. 19) are conspecific. They differ from the Brazilian material
only in having a slightly longer orifice and in having a rather coarsely granular
frontal wall.

Marcus (1938 : 233-234 key and distribution statements) working, as he explained,
from the literature without material for examination, described the ovicell area of
C. longiseta as without pores, thus distinguishing it from C. levinseni. He did,
however, refer the Mauritius form to C. longiseta, but also regarded the form from
the China Sea (C. soled) as synonymous.

Osburn (1940 : 452 ; 1952 : 479) thought that C. longiseta might be synonymous
with C. setigera (Smitt), but the ovicell and the orifice of the latter species are clearly
distinctive.

Distribution. Recent : Cuba (67-201 fathoms) (Canu & Bassler) ; John
Adams Bank, Brazil ; St. Helena (36-69 metres) (Marcus) ; Mauritius ; ? Zanzibar
(Waters) ; ? Providence, NNE. Madagascar (50-78 fathoms) (Thornely).

g. Crepidacantha grandis Canu & Bassler

(Fig. IG)

Crepidacantha grandis Canu & Bassler, 1929, Bull. U.S. nat. Mus. 100 (9) : 411, text-figs. 160L,
i6ia, pi. 57, figs. 4-7.

Syntypes. U.S. National Museum No. 8220. Recent : Philippines.

Material examined : —

B.M., 31. 12.30. 158 (Fig. ig). Recent: Albatross Stn. 5217, 105 fathoms,
Ragay Gulf, off North Burias, Philippines, in coarse, grey sand.

Diagnosis. Crepidacantha with large zooecia, the calcified ectooecium confined
almost entirely to the distal-lateral edges of the ovicell. Vibracula placed at the
level of the large opercular condyles, the flagella directed proximally. Proximal lip
of orifice very convex and rounded.

Remarks. This species appears to be quite distinct. Canu & Bassler state
(1929 : 412) that on the " nonmarginal zooecia the setiform spines [spines arising
from the marginal portions of the zooecia] are short (= 0-15 — 0-20 mm.) and all
equal ; they are very long on the contrary (= 0-60 mm.) on the free side of the
marginal zooecia."

In a number of zooecia in the British Museum specimen (B.M., 31 . 12.30. 158) the
frontal wall is punctured by a large rounded hole through which, presumably, the
polypide was extracted for food by some predatory borer.

Distribution. Recent : Philippines (105 fathoms).

h. Crepidacantha solea Canu & Bassler

(Fig. ih)

Lepralia poissonii Audouin (first form), Hincks, 1885, Ann. Mag. nat. Hist. (5) 15 : 256 (part.) ;
L. poissonii, Kirkpatrick, 1890 : 16.

ON THE POLYZOAN GENUS CRE P I D AC A N T H A LEVINSEN 253

Crepidacantha poissonii, Canu & Bassler, 19286, Bull. Soc. Sci. nat. med. Seine-et-Oise (2) 8 (7)
suppl. : 37, pi. 8, fig. 7 ; Canu & Bassler, 1930 : 33, pi. 5, fig. 5 ; C. poissoni, Osburn, 1952 :
478, pi. 58, fig. 2.

Crepidacantha solea Canu & Bassler, 1929, Bull. U.S. nat. Mus. 100 (9) : 409 ; Canu & Bassler,
1930 : 32.

Lectotype (here chosen). B.M., 89.8.21.19 (Fig. ih). Recent: Tizard Reef,
China Sea, 27 fathoms. Specimen encrusting lamellibranch. Labelled " Lepralia
poissonii Aud." by Kirkpatrick. H.M.S. " Rambler " Collection.

Other material examined : —

B.M., 97.5.1.828. Recent: Tahiti. Labelled "Lepralia Poissonii Aud."
by Miss Jelly. Bracebridge Wilson Collection.

B.M., 99.5.1.824. Recent: Tahiti. Labelled "Lepralia Poissonii Aud.
791. Avicularia horizontal " by Miss Jelly. Hincks Collection.

Note. Another slide, B.M., 99.5.1.823, Hincks Collection, similarly
labelled, has no specimen.

Diagnosis. Crepidacantha with a prominent transverse ridge across the
entooecium which also has a faint median carina and traces of longitudinal markings.
Calcified ectooecium confined to the distal-lateral margins. Vibracula large,
placed proximally to the orifice, the flagella directed laterally inwards. Proximal
lip of orifice square, plate-like.

Remarks. Canu & Bassler (1929 : 409) introduced the name Crepidacantha solea
for " Lepralia poissonni [sic], Kirkpatrick, 1888, from Mauritius, Sea of China and
Australia." Kirkpatrick's material from these localities is preserved in the British
Museum collections and proves to belong to three different species. Later (1930 : 32),
however, Canu & Bassler restricted the name C. solea to Kirkpatrick's material
from the China Sea (1890 : 16) and the species is here understood in that sense.

C. solea is, indeed, very closely related to the original Flustra poissonii Audouin,
especially in the appearance of faint longitudinal markings 1 on the frontal part of
the entooecium. It is distinguished by its large vibracula placed more proximally
than usual and by the square plate-like lip of the orifice.

Hincks (1885 : 256) recorded two forms of "Lepralia Poissonii, Audouin " from
Tahiti. The first, in which " ... the vibracula are situated below the orifice
and are placed horizontally," is undoubtedly the present species as shown by Hincks's
material (B.M., 97.5.1.828 and B.M., 99.5.1.824). The second (B.M., 99.5.1.825),
with vertical vibracula placed alongside the orifice, is probably C. crinispina
(Levinsen) .

It is evident that the material from the Hawaiian Islands and the Galapagos
Islands, identified by Canu & Bassler (1928& : 37 ; 1930 : 32, 33) as C. poissonii,
can also be placed here, as well as Osburn's material from the American Pacific
coast (1952 : 478, pi. 58, fig. 2).

Distribution. Recent : Tizard Bank, China Sea (27 fathoms) ; Tahiti ;
Hawaii (130-250 fathoms) (Canu & Bassler) ; Galapagos Islands (33-40 fathoms)
(Canu & Bassler) ; American Pacific coast (Osburn).

1 See Postscript.

254 ON THE POtYZOAN GENUS CREP I D AC A N T H A LEVINSEN

j. Crepidacantha zelanica Canu & Bassler

(Fig- iJ)

Lepralia poissonii Audouin (first form), Hincks, 1885, Ann. Mag. nat. Hist. (5) 15 : 256 (part.) ;

L. poissonii, Waters, 1887 : 59 (part.), pi. 8, fig. 37.
Crepidacantha poissoni (?) Waters, Canu & Bassler, 1929, Bull. U.S. nat. Mus. 100 (9) : 409,

text-fig. i6oj.
Crepidacantha zelanica Canu & Bassler, 1929, Bull. U.S. nat. Mus. 100 (9) : 410 ; Canu & Bassler,

1930 : 32, 33 ; Brown, 1952 : 362, text-figs. 285, 286.

Neotype (chosen by Brown, 1952:362). B.M., 99.5.1.826. Recent: Napier,
New Zealand. Specimen encrusting pebble. Labelled " Lepralia Poissonii, Aud."
by Miss Jelly. Hincks Collection.
Other material examined : —

Recent : Waitarere Beach, Wellington, 25 fathoms. Specimen encrusting
concretion. Geology Department, Otago University.

Recent : Kaka Point, Southland, New Zealand. Specimen encrusting
Atrina zelandica (Gray). Geology Department, Otago University.

B.M., D. 32531. Nukumaruan [Middle Pliocene] : Waipukurau Gorge.
Encrusting specimen. Labelled "Lepralia Poissonii Aud." by Miss Jelly.
Hincks Collection.

B.M., D. 36901. Nukumaruan [Middle Pliocene] : Napier Harbour. En-
crusting specimen. Hincks Collection.

B.M., D.1430, D.i 43 1. Nukumaruan [Middle Pliocene] : Petane. Two
encrusting specimens. Labelled "Lepralia Poissonii Aud." by Miss Jelly.
Vine Collection.

B.M., D. 36897 (Fig. ij)-36900. Nukumaruan [Middle Pliocene] : Petane.
Four encrusting specimens. Blake Collection.

Nukumaruan [Middle Pliocene] : Waipukurau. Encrusting specimen.
Labelled "Lepralia Poissonii" by Waters. Waters Collection, Manchester
Museum.

Nukumaruan [Middle Pliocene] : Petane. Encrusting specimen. Labelled
"Lepralia Poissonii" by Waters. Waters Collection, Manchester Museum.

Diagnosis. Crepidacantha with uncalcified area of ectooecium occupying most
of the frontal surface of the globular ovicell, exposing coarsely perforate entooecium.
Heterozooecia large, prominent, placed proximally to the orifice, the flagella directed
inwards. Orifice with slightly convex proximal lip.

Remarks. This species has been fully discussed by Brown (1952 : 362). In the
Recent specimen from Kaka Point, Southland, the heterozooecia, although fairly
large in size, are still vibracular in nature and have pronounced setiform flagella.

The first form of "Lepralia Poissonii, Audouin " recorded from New Zealand
by Hincks (1885 : 256) is almost certainly this species though no material labelled
by Hincks is available.

Distribution. Recent : Napier ; Wellington ; Kaka Point, New Zealand.

Fossil : Nukumaruan [Middle Pliocene] : Petane ; Waipukurau Gorge ; Napier
Harbour.

ON THE POLYZOAN GENUS C RE P I D A C A N T H A LEVINSEN

255

k. Crepidacantha bracebridgei 1 sp. nov.

(Fig. 2)

Holotype. B.M., 88 . 11 . 14 . 224 (Fig. 2). Recent : Port Phillip Heads, Victoria.
Specimen encrusting lamellibranch shell. Slide labelled "Lepralia poissonii Audn."
by Kirkpatrick. Bracebridge Wilson Collection.

Other material examined : —

B.M., 88. 11. 14.434. Paratype. Recent: Port Phillip Heads. Encrust-
ing specimen. Labelled "Lepralia. Jan. 1887 " by Kirkpatrick. Brace-
bridge Wilson Collection.

B.M., 97.5.1.833. Paratype. Recent: Port Phillip Heads. Encrusting
specimen. Labelled "Lepralia setigera. 1883 " by J. Bracebridge Wilson.
Bracebridge Wilson Collection.

B.M., 88 . 11 . 14 . 116. Paratype. Recent : Port Phillip Heads. Encrusting

specimen. Labelled " Lepralia setigera MacG. 1884 " by J. Bracebridge

Wilson. Bracebridge Wilson Collection.

Diagnosis. Crepidacantha with the porous entooecium visible in a narrow, oval,

longitudinal fenestra in the ectooecium. Vibracula placed on the distal side of the

opercular condyles, the flagella very long and directed proximally. A prominent

sub-oral tubercle usually present. Orifice long and narrow with a convex proximal

lip.

O 0.5 mm.

Fig. 2. C. bracebridgei, sp. nov. Holotype, B.M., 88. 11 . 14.224.
Recent : Port Phillip Heads, Victoria.

Description. Zoarium encrusting.

Zooecia ovate or hexagonal, arranged quincuncially in radiating rows, distinctly
separated by deep furrows. Orifice elongate, divided into an oval anter and a shallow
poster, with a convex or squared proximal lip, the opercular condyles strong.
Peristome thin and raised distally into a hood on non-ovicelled zooecia. Frontal

1 After Mr. J. Bracebridge Wilson, the Australian naturalist.

256 ON THE POLYZOAN GENUS C RE P I D A C A N T H A LEVINSEN

wall ventricose, finely granular and porous, with a prominent sub-oral tubercle.
Marginal spines about twelve in number.

Vibracula paired, placed above the level of the opercular condyles, the flagella
directed proximally and reaching beyond the proximal margins of the zooecium.

Ovicells globular, hyperstomial, longer than wide, with the uncalciried area of the
ectooecium small, longitudinally-oval, placed towards the distal end of the ovicell,
somewhat raised. Often a distinct umbo at its proximal end. Exposed entooecium
porous.

Measurements. L z = 0-50-0-55 mm. ; l z == 0-29-0-33 mm.
h r = 0-08-0-09 mm - '> K = 0-07-0-08 mm.

Remarks. The ovicell of this species is very distinctive and differs from that of
C. longiseta in that the porous area is always raised and not depressed or surrounded
by salient walls as it is in the latter species. The area is, moreover, usually placed
transversely in C. longiseta. The presence of a sub-oral tubercle in the Australian
species is probably of no great significance as its size appears to vary with the age
of the zooecia, but the hooded peristome of the non-ovicelled zooecia may be an
important character.

Although two of the specimens mentioned above in the Bracebridge Wilson
Collection (B.M., 97.5.1.833 and B.M., 88. n. 14. 116) are labelled " Lepralia
setigera," they do not belong to the species described by MacGillivray as Lepralia
setigera Smitt (1883 : 133, pi. 1, figs. 2, 3) as his description and figure of the ovicell
are clearly those of C. crinispina Levinsen sp.

Distribution. Recent : Port Phillip Heads, Victoria.

l. Crepidacantha kirkpatricki 1 sp. nov.

(Fig- 3)
Lepralia poissonii Audouin, Kirkpatrick, iSS8b, Ann. Mag. nat. Hist. (6) 2 : 14.
Lepralia poissonii Audouin var. Waters, 1889, J. R. micr. Soc. : 14, pi. 2, fig. 17.
Crepidacantha poissoni Audouin var. Levinsen, 1909, Morph. Syst. Stud. Cheil. Bry. : 268,

text-fig. 4 ; Canu & Bassler, 1923 : 174, text-fig. 33D.
Crepidacantha crinispina Levinsen var. Canu & Bassler, 1929, Bull. U.S. nat. Mus. 100 (9) : 410,

text-fig. 160G.
Crepidacantha levinseni Marcus, 1938, Vidensk.Medd.dansk naturh.Foren. Kbh. 101 : 231 (part. —

non text-figs. 2 8a, b).

Holotype. B.M., 88. 11. 14. 426 (Fig. 3). Recent : Port Phillip Heads, Victoria.
Specimen encrusting Fucus. Labelled " Lepralia poissonii Audn." by Kirkpatrick
and " Jan. 1887 " by J. Bracebridge Wilson. Bracebridge Wilson Collection.
Other material examined : —

B.M., 97.5.1.834. Paratype. Recent: Port Phillip Heads. Labelled
"Lepralia setigera. 1885 " by J. Bracebridge Wilson. Bracebridge Wilson
Collection.

Paratype : Universitetets Zoologiske Museum, Copenhagen No. 1/8/90.
Recent : Port Phillip Heads. Levinsen Collection.

1 After the late Dr. R. Kirkpatrick, formerly of the Department of Zoology, British Museum (Natural
History).

ON THE POLYZOAN GENUS C RE P I D A C A NT H A LEVINSEN

257

Diagnosis. Crepidacantha with the ovicell flattened and deeply immersed in
the distal zooecium, the uncalcified area of the ectooecium occupying almost the
whole of the exposed frontal surface. Vibracula placed proximally to the orifice,
the flagella short and directed laterally inwards. A small sub-oral tubercle present.
Proximal lip of orifice squared, plate-like. Marginal spines of mature zooecia short
and thick, of marginal zooecia long and flexible. 1

Description. Zoarium encrusting.

Zooecia ovate, arranged somewhat irregularly in radiating rows, scarcely separated
by shallow furrows. Orifice sub-circular distally with a shallow poster separated
by inconspicuous condyles, the proximal lip square and plate-like. Peristome thin,
scarcely raised. Frontal wall usually flattened, occasionally ventricose, finely
perforate and covered with small papillae, a small, sub-oral tubercle often present.
Marginal spines of mature zooecia short and thick, eight in number, those of the
young zooecia long and flexible.

j 1 1 1

O 0.5mm.

Fig. 3. C. kirkpatricki, sp. nov. Holotype, B.M., 88. 11 . 14.426.
Recent : Port Phillip Heads, Victoria.

Vibracula paired, placed proximally to and on each side of the orifice, the short
flagella directed obliquely laterally inwards.

Ovicells large, sub-circular in outline, deeply immersed in mature zooecia, the
exposed frontal surface flattened and occupied almost entirely by a porous entooecium,
the calcified ectooecium marginal.

Measurements. L z = 0-35-0-40 mm. ; l z = 0-25-0-28 mm.
h r = 0-09-0-10 mm. ; l r = 0-07-0-08 mm.

Remarks. The holotype is part of the material described by Kirkpatrick
(1888& : 14) as Lepralia Poissonii, Audouin. I have also examined the specimen
described by Levinsen (1909 : 268) as Crepidacantha poissoni var. , which was kindly
lent by Dr. P. L. Kramp of Universitetets Zoologiske Museum, Copenhagen, and
this proves to be a small fragment, also encrusting Fucus, which was sent to Levinsen

1 See Postscript for further information about this species.

258 ON THE POLYZOAN GENUS CREPI D A C A NT H A LEVINSEN

by Miss Jelly in 1890. It is unquestionably from the same material, if not from the
same colony, as the holotype.

This is also the species described by Waters (1889 : 14) from Green Point, New
South Wales, in which the ovicells are " immersed, showing, however, a round
ovicellular area on a level with the wall of the zooecium." Canu & Bassler (1929 :
410) thought that this probably belonged to another genus.

This species is closely related to C. solea Canu & Bassler, but differs from it in
having a large, flattened ovicell and eight stout marginal spines, the spines of C.
solea being slender and 12 to 14 in number. In C. kirkpatricki the surface of the
zoarium is fairly even and the zooecia are not usually separated by furrows but are
indistinctly merged with one another. The proximal lip of the orifice is a distinct
median plate inclined to turn upwards and outwards from the orifice, the lateral
indentations appearing as small sub-circular openings.

Marcus (1938 : 231), in proposing the new name C. levinseni for Levinsen's variety
of C. poissonii (1909 : 268) and for his own material from St. Helena, stated that he
had not seen the Port Phillip Heads material. As shown above, the St. Helena
species is most probably C. longiseta Canu & Bassler.

Distribution. Recent : Port Phillip Heads, Victoria ; Green Point, Port
Jackson, N.S.W. (Waters).

m. Crepidacantha odontostotna (Reuss)

Lepralia odontostoma Reuss, 1874, Denkschr. Akad. Wiss. Wien. 33 : 156, pi. 4, fig. 8.
Crepidacantha (Lepralia) odontostoma, Canu & Bassler, 1930, Proc. U.S. nat. Mus. 76 (13) : 32.

Remarks. Without seeing Reuss's material, which comes from the Miocene of
" Rauchstallbrunngraben bei Baden," it is difficult to decide on the generic affinities
of this species. The orifice has a straight or slightly concave proximal lip and the
paired heterozooecia are placed at the level of the opercular condyles. If this is a
species of Crepidacantha and if the preservation is as good as Reuss's figure suggests,
then one would expect to see some traces of marginal areolae. The absence of
ovicells makes identification even more difficult.

doubtful references to species of Crepidacantha

Thornely has recorded Lepralia poissonii Audouin from Ceylon (1905 : 119) and
the Andamans 1 (1907 : 190). It is impossible to identify the species, but, from
comparison with the world distribution map (Fig. 4), it is possible that C. longiseta
Canu & Bassler or C. crini spina Levinsen sp. is represented here.

Waters (1914 : 832, 856) recorded Lepralia poissonii Audouin from the Atlantic
[? = C. poissonii var. teres], Indian Ocean [? = Thornely 's record, probably C.
longiseta], Australia, and Japan. I did not see Waters's material from these
localities in the collections at Manchester Museum, and, therefore, I cannot suggest
to which species the Australian and Japanese records refer.

1 A specimen (B.M., 84.3.25.1, Wilmer Collection) from the Andaman Islands, labelled "Lepralia
poissonii Aud." by Kirkpatrick, is not a Crepidacantha.

ON THE POLYZOAN GENUS CRE P I DA C A NT H A LEVINSEN 259

Fig. 4. Recent distribution of species of Crepidacantha. Species are indicated by the
index letters used throughout the text.

2 6o ON THE POLYZOAN GENUS CREP I D AC AN T H A LEVINSEN

CONCLUSIONS

The known species of Crepidacantha Levinsen are n or 12 in number and, until
recently, had been recorded only from tropical and sub-tropical waters. They have
recently been found, however, as far south as Kaka Point, Southland, New Zealand
(169 50' E., 46 25' S.), on a coast-line swept by a northerly drift. This i9 in
accordance with their presence in rocks of Nukumaruan [Middle Pliocene] age in
Hawkes Bay. This region (about 40 S.) has been shown by Fleming (1944) from
molluscan evidence to be the northern limit of subantarctic waters in Nukumaruan
times. Species of Crepidacantha are also plentiful in the Waiauan [Middle Miocene]
rocks of North Canterbury, New Zealand, rocks deposited when New Zealand was
experiencing a tropical climate.

In the northern hemisphere Crepidacantha has not been recorded north of Madeira
(33° N.), though Waters's record from Japan (1914 : 832), if correct, may extend
this range.

The first species described, Flustra poissonii Audouin, is not at all cosmopolitan
in distribution, and, in fact, appears to be restricted to the Eastern Mediterranean
or Red Sea area. 1 The accompanying map (Fig. 4) shows the known distribution
of the various species.

POSTSCRIPT 2

Dr. Anna B. Hastings, in litt., writes :

" In his forthcoming, posthumous 'Siboga ' Report, Sir Sidney Harmer remarks
that a revision of the genus Crepidacantha is urgently needed.

" Pending such a revision he has recognized only two species in the Indo-Malayan,
region — C. poissonii and C. crinispina — and has given tentative synonymies
referring much of the previously recorded material to one or other of these species.
You have now provided the necessary revision discriminating more species in this
material and superseding (although ante-dating) Harmer 's synonymies, which, but
for the war of 1939-1945, might have been published ten years ago.

" Certain material which, as you mention, was not available when you did your
work was considered by Harmer and has since been presented by him to the British
Museum (Natural History). I have examined the specimens and you may like to
include some notes about them :

" 1. Two specimens from Ghardaqa on the Egyptian coast of the Red Sea
(Crossland, 4.P on slides 4.G 1 and 4.G 3 of Arachnopusia spathulata). The material
' agrees closely with Savigny's fig. 5 2 (PL X) except in not showing radial markings
on the ovicell ' (Harmer MS.).

" In the absence of any record of Crepidacantha from the Mediterranean, the
presence on the Red Sea coast of Egypt of a species agreeing closely with Savigny's
figure strongly suggests that the material of Savigny & Audouin came from the
Red Sea. I am prepared to accept the two specimens from Ghardaqa as repre-
senting true C. poissonii.

1 See Postscript.

2 The information given in this Postscript is incorporated in the distribution-map.

ON THE POLYZOAN GENUS CREP I D A C ANT H A LEVINSEN 261

" C. poissonii (as represented by the Ghardaqa specimens) and C. solea are, as
you have suggested, very closely related. They may perhaps be distinguished by
the more proximal position of the heterozooecia in C. solea (which are directed
laterally inwards in both species), but the Ghardaqa material shows some variation
in this character. The ovicells of both show the proximo-lateral calcareous layer
described by you in C. poissonii var. teres, where it is sometimes conspicuous for
its opaque whiteness. They also agree with that variety in having a porous,
transverse ridge on the ovicell. The pores are, however, more or less obscured by a
membranous layer (membranous part of the ectooecium). The appearance depends
very much on the direction of the light. Illuminated from the proximal end the
membrane covering the frontal surface shines and, being stretched in drying, may
give an impression of some radial striation, and the pores are hidden. Illuminated
from the distal end the membranous layer is inconspicuous and the transverse row
of pores may be observed. Such effects as this probably account for the absence
of any appearance of the row of pores on the entooecium in your figures of C. crini-
spina var. parvipora both here (Fig. id) and in the Catalogue (Brown, 1952 : 361,
text-fig. 284) and of typical C. crinispina (Catalogue, text-fig. 283).

" 2. Thornely's Ceylon specimen (Thornely Coll., 1906, no. 46). This specimen
agrees very closely with the Ghardaqa specimens of C. poissonii. It has one ovicell,
which shows the membrane, the row of pores, and the proximo-lateral calcareous
ectooecium as described above.

" 3. Thornely's Andaman specimen (Thornely Coll., 1935, no. 118). This belongs
to C. crinispina as limited by you.

" 4. A specimen from Victoria (Cambridge Museum, E. C. Jelly, May 24, 1895) is
also of interest. Harmer (MS.) mentions it as probably representing a new species.
It has no ovicells, but in other respects agrees with the type-specimen of C.
kirkpatricki. Two parallel, longitudinal, oval bosses on the opercula, mentioned by
Harmer, are also present in your type-specimen. They are blister-like prominences
on the outer surface of the operculum, and are conspicuous when the specimens
(which are dry) are examined under a binocular microscope."

REFERENCES

Audouin, V., & Savigny, M. J. C. L. de. 1826. Description de VEgypte. Histoire Naturelle,
I , 4, Explication sommaire des Planches dont les dessins ont ete fournis par J.-C. Savigny,
pp. 225-244, pis. 1-14. Paris.

— — ■ — ■ — ■ 1828. Description de VEgypte. 2nd. Ed., XXIII, Histoire Naturelle, Zoologie,
Explication sommaire des Planches de Polypes de VEgypte et de la Syrie, Publiees par Jules-
Cesar Savigny. ... , pp. 40-78. Paris.

Bassler, R. S. 1935. Fossilium Catalogus I : Animalia, 67 : Bryozoa. 's-Gravenhage.

Brown, D. A. 1952. The Tertiary Cheilostomatous Polyzoa of New Zealand. Cat. Brit. Mus.
(Nat. Hist.), pp. 1-405.

— — 1954- Polyzoa from a Submerged Limestone off the Three Kings Islands, New Zealand.
Ann. Mag. nat. Hist. (12) 7: 415-437.

Busk, G. 1854. Catalogue of Marine Polyzoa in the British Museum, Part II, pp. 55-120,
pis. 69-124. London.

262 ON THE POLYZOAN GENUS CRE P I D A C AN T H A LEVINSEN

Calvet, L. 1907. Bryozoaires in Exped. Sci. " Travailleur " et " Talisman," pp. 355-495,

pis. 26-30. Paris.
Canu, F., & Bassler, R. S. 191 7. A Synopsis of American Early Tertiary Cheilostome Bryozoa.

Bull. U.S. nat. Mus. 96 : 1-87, pis. 1-6.

1920. North American Early Tertiary Bryozoa. Ibid., 106: i-xx, 1-879, pis. 1-162.

— ■ — — — 1923. North American Later Tertiary and Quaternary Bryozoa. Ibid. 125 : i-vii,

1-302, pis. 1-47.
— ■ — — — ■ 1927. Classification of the Cheilostomatous Bryozoa. Proc. U.S. nat. Mus. 69

(14) : 1-42, pi. 1.
— '1928a. Fossil and Recent Bryozoa of the Gulf of Mexico Region. Ibid. 72 (14) :

1-199, pis. 1-34.
— ■ — ■ 19286. Bryozoaires des lies Hawaii. Bull. Soc. Sci. nat. med. Seine-et-Oise, (2) 8,

7, suppl. : 33-66, pis. 6-1 1.
■ — 1929. Bryozoa of the Philippine Region. Bull. U.S. nat. Mus. 100 (9) : 1-685, V^ s -

1-94.

— — 1930. The Bryozoan Fauna of the Galapagos Islands. Proc. U.S. nat. Mus. 76

(13) : 1-78, pis. 1-14.

& Lecointre, G. 1930. Les Bryozoaires cheilostomes des faluns de Touraine et

d'Anjou. Mem. Soc. geol. Fr. (n.s.) 4 : 83-130, pis. 16-25.
Edwards, F. W. 1910. Bryozoa in Zool. Rec. (1909) 46 : 14-26.
Fleming, C. A. 1944. Molluscan Evidence of Pliocene Climatic Change in New Zealand.

Trans, roy. Soc. N.Z. 74 : 207-220.
Hincks, T. 1880. Contributions towards a General History of the Marine Polyzoa. Part I.

Ann. Mag. nat. Hist. (5) 6 : 69-92, pis. 9-1 1.
— — ■ 1881. Contributions towards a General History of the Marine Polyzoa. Ibid. (5) 8 : 122-

136, pis. 1-5.
— — 1885. Contributions towards a General History of the Marine Polyzoa. Ibid. (5) 15 : 244-

257, pis. 7-9.
— > — ■ 1891. Contributions towards a General History of the Marine Polyzoa. — Appendix.

Ibid. (6) 8 : 86-93.
■ 1895. Index to " Marine Polyzoa : Contributions towards a General History," pp. i-vi.

London.
Jullien, J., & Calvet, L. 1903. Bryozoaires provenant des campagnes de YHirondelle.

Result. Camp. sci. Monaco, 23 : 1-188, pis. 1-18.
Kirkpatrick, R. 1888a. Polyzoa of Mauritius. Ann. Mag. nat. Hist. (6) 1 : 72-85, pis.

7-10.

- 18886. Polyzoa from Port Phillip. Ibid. (6) 2 : 12-21, pi. 2.

1890. Report upon the Hydrozoa and Polyzoa collected ... by H.M.S. " Rambler."

Ibid. (6) 5 : 11-24, P* s - 3 _ 5-

Levinsen, G. M. R. 1909. Morphological and Systematic Studies on the Cheilostomatous

Bryozoa, vii -f 431 pp., 27 pis. Copenhagen.
Livingstone, A. A. 1929. Bryozoa Cheilostomata from New Zealand. Vidensk. Medd.

dansk naturh. Foren. Kbh. 87 : 45-104, pis. 1, 2.
MacGillivray, P. H. 1883. Descriptions of New, or Little-Known, Polyzoa. Part II. Trans.

roy. Soc. Vict. 19 : 130-138, pis. 1-3.
— ■ — • 1887. A Catalogue of the Marine Polyzoa of Victoria. Ibid. 23 : 187-224.
Marcus, E. 1938. Bryozoen von St. Helena. Vidensk. Medd. dansk naturh. Foren. Kbh.

101 : 183-252.
Norman, A. M. 1909. The Polyzoa of Madeira and neighbouring Islands. /. Linn. Soc.

Lond. {Zool.) 30 : 275-314, pis. 33-42.
d'Orbigny, A. D. 1851-1854. Paleontologie Francaise. Terrains Cretaces. 5. Bryozoaires :

1-1192, pis. 600-800. Paris.

ON THE POLYZOAN GENUS CREPI D AC ANTH A LEVINSEN 263

Osburn, R. C. 1940. Bryozoa of Porto Rico with a Resume of the West Indian Bryozoan
Fauna. Sci. Surv. Porto Rico and Virgin Islands, 16 (3) : 321-486, pis. 1-9.

■ 1952. Bryozoa of the Pacific Coast of America. Part 2. Cheilostomata — Ascophora.

Allan Hancock Pacif. Exped. 14 (2) : 269-611, pis. 30-64.

Philipps, E. G. 1899. Report on the Polyzoa collected by Dr. Willey from the Loyalty
Isles, New Guinea and New Britain. A. Willey' s Zoological Results, 4 : 439-450, pis. 42,

43-
Reuss, A. E. 1874. Die fossilen Bryozoen des osterreichisch-ungarischen Miocans. I.

Abtheilung. Denkschr. Akad. Wiss. Wien, 33 : 141-190, pis. 1-12.
Smitt, F. A. 1873. Floridan Bryozoa. Part II. K. svenska Vetensk.-Akad. Handl. 11 (4) :

1-83, pis. 1-13.
Thornely, L. R. 1905. Report on the Polyzoa . . . Rept. Govt. Ceylon Pearl Oyster

Fisheries Gulf Manaar, Roy. Soc. London, 4 Suppl. Rept. 26 : 107-130, pi. 1.

1907. Report on the Marine Polyzoa in the collection of the Indian Museum. Rec.

Indian Mus. 1 : 179-196.

1912. The Marine Polyzoa of the Indian Ocean, from H.M.S. " Sealark." Trans. Linn.

Soc. Lond. (Zool.) (2) 15 : 137-157, pi. 8.

Vigneaux, M. 1949. Revision des Bryozoaires neogenes du Bassin d'Aquitaine et essai de

classification. Mem. Soc. geol. Fr. 60 : 1-155, pis. 1-11.
Waters, A. W. 1887. On Tertiary Chilostomatous Bryozoa from New Zealand. Quart. J.

geol. Soc. Lond. 43 : 40-72, pis. 6-8.

1889. Bryozoa from New South Wales. Ann. Mag. nat. Hist. (6) 4 : 1-24, pis. 1-3.

— — ■ 1899. Bryozoa from Madeira. /. R. micr. Soc. 1899 : 6-16, pi. 3.

1 91 4. The Marine Fauna of British East Africa and Zanzibar, . . . Bryozoa— Cyclo-

stomata, Ctenostomata and Endoprocta. Proc. zool. Soc. Lond. 1914 : 831-858, pis. 1-4.

PRESENTED
8 NOV 1954

PRINTED IN GREAT BRITAIN BY
ADLARD AND SON, LIMITED,
BARTHOLOMEW PRESS, DORKING

•sENT^j BDIV lSIONS OF THE

3 NOV 1954

GENUS PTILINOPUS

(AVES, COLUMBAE)

A. J. CAIN

BULLETIN OF
THE BRITISH MUSEUM (NATURAL HISTORY)
ZOOLOGY Vol. 2 No. 8

LONDON: 1954

SUBDIVISIONS OF THE GENUS PTILINOPUS

(AVES, COLUMBAE)

ARTHUR JAMES CAIN j

(Department of Zoology and Comparative Anatomy, University Museum, Oxford)

Pp. 265-284 ; 2 Plates

BULLETIN OF

THE BRITISH MUSEUM (NATURAL HISTORY)

ZOOLOGY Vol. 2 No. 8

LONDON: 1954

THE BULLETIN OF THE BRITISH MUSEUM
(NATURAL HISTORY), instituted in 1949, is
issued in five series corresponding to the Departments
of the Museum, and an Historical Series.

Parts appear at irregular intervals as they become
ready. Volumes will contain about three or four
hundred pages, and will not necessarily be completed
within one calendar year.

This paper is Vol. 2, No. 8 of the Zoological series.

PRINTED BY ORDER OF THE TRUSTEES OF
THE BRITISH MUSEUM

Issued October, 1954 Price Eight Shillings

SUBDIVISIONS OF THE GENUS PTILINOPUS

(AVES, COLUMBAE)
By ARTHUR JAMES CAIN

CONTENTS

Page
Introduction ........... 267

Principles ........... 268

(1) Recognition of natural groups . . . . . . .268

(2) Size 269

(3) Keys and natural classifications . . . . . .269

(4) Use of species-groups and subgenera . . . . .270

(5) The superspecies . . . . . . . . .270

Classification ........... 271

(1) Brief list 272

(2) Detailed classification . . . . . . . .273

Notes on the Classification . . . . . . . 278

(1) Generic and subgeneric limits . . . . . . .278

(2) Leucotreron and Ramphiculus . . . . . . .280

(3) Ptilinopus . . . . . . . . . .281

(4) The purpuratus species-group . . .... 282

Summary 284

References ........... 284

SYNOPSIS

Previous arrangements of the species of Ptilinopus result in unnatural groupings. When one
abandons the attempt to provide diagnoses for all groups, and gives due weight to the facts
of geographical distribution, a far more natural classification can be constructed.

INTRODUCTION

The genus Ptilinopus contains a large number of brightly coloured fruit-pigeons
distributed from the Malay Peninsula, Philippines and Marianas to New South
Wales, the Marquesas, and Henderson Island. Even within obviously natural
groups of species of this genus so much interspecific variation occurs that it is almost
impossible to find diagnostic characters. The last comprehensive review with
diagnoses is that by Salvadori (1893) which, as was usual at that period, ascribes
exaggerated importance to so-called anatomical characters. Some species have
been reviewed by Rensch (1929), Mayr (1940), Ripley and Birckhead (1942) and
Amadon (1943), but no general survey has appeared, except that Peters (1938)
reviewed the limits of the genera of fruit-pigeons and gave diagnoses for those

ZOOL. II, 8, i6§

2 68 SUBDIVISIONS OF THE GENUS PTILINOPUS

genera he recognized, together with a very brief survey of the interrelations of
species within each genus. The checklist given by Peters (1937) is good for sub-
specific ranges and synonymy, but the arrangement of the species is unfortunate,
several closely related forms being separated widely. Also, no indication is given
of the limits of superspecies or species-groups.

The purpose of this paper is to present a classification of these pigeons, based on
a study of the collections in the British Museum (Natural History). I am deeply
indebted to Mr. J. D. Macdonald, who has given me every facility in examining the
specimens, and to Mr. R. E. Moreau for advice and criticism.

PRINCIPLES
(1) Recognition of natural groups

If each of the members of a group of species is more closely related to some other
members than to any species outside it, the group is natural ; but it may well
happen that every character that is confined to the group and therefore might be
proposed as diagnostic is lost or modified in one or more species. If those species
are clearly related by all their other characters to the rest of the group, they cannot
be excluded from it.

This is the state of affairs found in the genus Ptilinopus. Consequently, the
descriptions given below must be understood as indicating general characteristics,
not as rigidly defining each subdivision. As Peters (1938) remarked of the subgenus
Ptilinopus, it is very difficult to draw up a diagnosis that applies to all the species.
I would add that it is unnecessary. When a natural group can be recognized but
not diagnosed as such, it is sufficiently indicated by a description and a list of the
contained species, as in the present paper.

When two species are clearly very closely allied but differ in one conspicuous
character, it is obvious that in respect to these two species this character is only of
specific importance. In another group, its distribution may coincide with that of
so many other characters that it can be cited as characteristic of a genus, family, or
group of even higher rank. The test of taxonomic importance of any character in
a particular group is its distribution in that group with respect to all other characters.

This statement is true whatever the nature of the character — anatomical, physio-
logical, embryological, genetical or ecological. The remark of Manuel (1936&) that
" The subfamily Ptilopodinae is an artificial group for the reason that there are no
trenchant structural characters peculiar to it " is consequently unacceptable. In
some natural groups of birds the beak, feet, or structure of the feathers may show
extraordinary variability while the colour pattern remains very constant ; in others,
the reverse is true. Consequently there is no need to assume a priori that in any
group of birds structure is always more important taxonomically than colour-
pattern. This was very nearly the assumption made by Salvadori (1893) as of
course by most authors of that time. Consequently his keys almost always begin
with an " anatomical " character (emargination of the first primary, length of the
tail and wing, or strength of the beak). Unfortunately these characters are not
invariably of the highest importance in Ptilinopus and its allies, and Peters (1938),

SUBDIVISIONS OF THE GENUS PTILINOPUS 269

although he implies (p. 376) that external structural characters are a sine qua non
for erecting a subfamily (which may be true in practice but cannot be made into an
absolute rule), remarks of Ptilinopus and its allies that " On the whole, colour in
this group seems just as important, if not more so than structure." A survey of all
the species, originally made without reference to Peter's paper, confirms his opinion.

(2) Size

The importance of any character, as was shown above, is determined only by its
distribution in any particular group with respect to all other characters, and no one
class of characters can be regarded on principle as always of greater taxonomic
importance than any other.

Equally, no one class of characters can be considered on principle as always
useless. Size tends to vary greatly in birds both between and within species, and
is therefore not usually of much taxonomic value in groups above the level of the
species. However, in the genus Ptilinopus size tends to remain comparatively
constant in the species of some groups. The subgenera Leucotreron and Ramphi-
culus contain species which, for the genus, are rather or very large (wing-length
150-170 mm.). In the subgenus Ptilinopus all the members of the ornatus species-
group ((io)-(i4)) are rather large for this subgenus (wing-length 135-160 mm.),
while those of the purpuratus species-group ((i5)-(3o)) with a solitary exception
(Pt. huttoni (26)) are rather small (wing-length 95-145 mm.). Size is therefore
included as a character in the descriptions of the groups. Within the genus as a
whole wing-lengths of 90-139 mm. are considered small, 140-159 medium, and
160-170 large.

(3) Keys and natural classifications

Keys are constructed for the rapid identification of specimens. The characters
employed are therefore those that are readily discernible, present in the largest
possible number of individuals (not confined, for example, to a fleeting stage of
development), least variable in their expression, and most easily defined verbally.

" Natural " classifications, on the other hand are constructed to display the
general affinities and presumed evolutionary interrelationships of different forms.
For this purpose characters totally different from those used in a key to the same
group may be chosen, since the conspicuousness of a character bears no direct
relationship to its taxonomic and phylogenetic importance. Yet monographs are
still often burdened with " keys " which are designed to act both as guides to
identification and as succinct accounts of a natural classification. In such " keys "
it is common to find not only that the characters employed are often of little practical
use for identification, but also that the necessity for defining each group by one or
two characters may actually prevent the recognition of such natural groups as
those described immediately above, which have no diagnostic characters. The
reason why such " keys " are constructed is, of course, that by using them a new
and undescribed form can usually be put into its natural group, whereas in a key
made solely for identification it may very well be associated with widely differing

270

SUBDIVISIONS OF THE GENUS PTILINOPUS

species with which it happens to have some unimportant but conspicuous character
in common. (For example, the character " bifid breast feathers " would bring
together Pt. porphyrea (s.g. Leucotreron) and Ducula goliath, and most of the
purpuratus species-group of Ptilinopus.) Nevertheless, the practice of combining
both functions in the same key is not good. Together with the assumption that
structure is always more important than colour, it was responsible for some of
Salvadori's least successful groupings in Ptilinopus, as can be readily verified by
anyone who will work through his classification. However, it is a striking tribute
to Salvadori's genius that although he worked under these unnecessary limitations,
his classification was by far the best produced until then, and most of his groups
can still be regarded as natural.

(4) Use of species-groups and subgenera

Mayr (1942 : 290) has pointed out the great taxonomic advantages of the species-
group. It is an informal natural group of closely related species, below the rank of a
subgenus, the recognition of which involves no complication of the scientific names
of the species contained in it. Moreover, having (unlike the subgenus) no formal
name, it is not readily raised to the rank of a genus with a consequent change of
the generic name throughout its component forms. In addition, in Ptilinopus, its
use permits the recognition of natural groups of species without a general promotion
of the subgenera to genera, and the genus perhaps to a subfamily. By using
species-groups, the generic name can be retained throughout (which indicates the
close relationship of the included species) while unmanageably large groups of
species can nevertheless be broken up.

As the species-group is burdened by no formal rules there is no necessity to use
the oldest valid name within a group as the name of that group. Consequently the
group containing as its oldest named form the very atypical Pt. tannensis (14)
can be called the ornatus group after a much more typical member (12). I have not
changed the name of the purpuratus species-group (Ripley & Birckhead, 1942)
because one should avoid nomenclatural disturbances wherever possible, and
Pt. purpuratus (27), although a very simplified form, is certainly a member of
the species-group that bears its name.

Although the subgenus has the disadvantage that it lengthens the names of its
component forms in full citations, and is a standing temptation to reviewers to
raise it to generic rank, it is nevertheless useful when, as in the genus Ptilinopus,
many species-groups must be recognized but they fall clearly into two or more
groups within the genus. Consequently it is used here to distinguish the three
very natural groups which the genus Ptilinopus comprises. In smaller or less
heterogeneous genera the species-group should be sufficient, without the use of
subgenera.

(5) The superspecies

The superspecies (Mayr, 1942 : 169) is a monophyletic group of forms which
replace each other geographically but are too diverse for all to be ranked as sub-

SUBDIVISIONS OF THE GENUS PTILINOPUS 271

species of a single species. The limit of permissible diversity cannot of course be
specified, since in some groups good species are very similar, in others extremely
diverse, so that in the former the diversity permissible among subspecies of a single
species will be much less than in the latter, and forms with striking characters will
be unhesitatingly separated as distinct species. Mayr (1942) gives examples of
monophyletic arrays of geographical representatives in which some forms have
even been segregated into a separate genus. Consequently a single superspecies
may include only a few species of a single species-group, as in the occipitalis species-
group ((4)-(7)), or all of them as in the lechlancheri species-group ((8)-(o,)), or all of
one species-group and one or more of another, as, for example, in the genus Halcyon
(Mayr, 1942).

Consequently, the superspecies cannot strictly be used as a rank in the taxonomic
hierarchy. However, when a series of geographically replacing forms is very
heterogeneous there is usually some doubt as to whether it is really monophyletic,
and it would be unwise, therefore, to list it as a superspecies. If there is no doubt,
then most workers would consider that, since geographical variation is considerable
in this series, the limits of subgeneric and generic variation in the natural group
containing it should be wide. Consequently it would be given a low rank in the
hierarchy. As a result it is rare to find a superspecies that transgresses the limits
of a species-group, and in practice the superspecies can be regarded as a rank
between the species and the species-group. The hierarchy of ranks used in the
following classification is therefore :

Genus,
Subgenus,

Species-group,

Subgroup (of species-group),
Superspecies,
Species.

CLASSIFICATION

The proposed classification is illustrated by the diagrams (Figs. 1 and 2) in which
are represented all the species recognized by me. These diagrams are intended to
demonstrate only the plumage pattern of each species. Relative size is not given
and in each species the most highly ornamented subspecies and sex are shown
(except in Pt. rivoli (32) and solomonensis (31) where there is very confusing con-
vergence) . Each species is given the same number in the diagrams as in the classi-
fication. Cross-hatching is used to represent shades of red, from orange to purple.
All other colours are represented by stippling of appropriate darkness. Except for
Pt. merrilli (5), specimens of all species and nearly all subspecies have been examined.
The diagram of Pt. merrilli is based on the coloured plate of P. m. faustinoi given by
Manuel (19360), and on the descriptions by Delacour and Mayr (1946) and McGregor
(1916). Full references to original descriptions of all forms will be found in Peters
(1937) and Salvadori (1893).

272

SUBDIVISIONS OF THE GENUS PTILINOPUS

(1) Brief list

Genus Ptilinopus

Subgenus Leucotreron

Subgenus Ramphiculus

(a) Occipitalis species-group
(i) Marchei subgroup

(ii) Occipitalis subgroup

(b) Leclancheri species-group

Subgenus Ptilinopus

(a) Ornatus species-group
(i) Wallacii subgroup

(ii) Ornatus subgroup

(iii) Tannensis subgroup

(b) Purpuratus species-group
(i) Superbus subgroup

(ii) Purpuratus subgroup

Pt. cincta superspecies
Pt. porphyrea (1)
Pt. dohertyi (2)
Pt. cincta (3)

Pt. marchei (4)
Pt. merrilli (5)

Pt. occipitalis superspecies
Pt. occipitalis (6)
Pt.fischeri (7)

Pt. leclancheri superspecies
Pt. leclancheri (8)
Pt. subgularis (9)

Pt. wallacii (10)

Pt. aurantiifrons (n)

Pt. ornatus (12)
Pt. perlatus (13)

Pt. tannensis (14)

Pt. superbus superspecies
Pt. superbus (15)
Pt. perousii (16)

Pt. monacha (17)
Pt. coronulatus (18)
Pt. pulchellus (19)
Pt. regina (20)
Pt. roseicapilla (21)
Pt. greyii (22)
Pt. richardsii (23)
Pt. porphyraceus (24)

SUBDIVISIONS OF THE GENUS PTILINOPUS

273

(ii) Viridis subgroup

(d) Hyogastra species-group
(i) Iozonus subgroup

(ii) Hyogastra subgroup

Pt. rarotongensis (25)
Pt. huttoni (26)
Pt. purpuratus (27)
Pt. insularis (28)
Pt. mercieri (29)
Pt. dupetithouarsii (30)

Pt. solomonensis (31)
Pt. rivoli (32)

Pt. viridis (33)

Pt. iozonus superspecies
Pt. iozonus (34)
Pt. insolitus (35)

Pt. hyogastra superspecies
Pt. hyogastra (36)
Pt. granulifrons (37)
Pt. naina (38)
Dubious member of this super-
species
Pt. melanospila (39)

Pt. jambu (40)

Pt. luteovirens superspecies
Pt. victor (41)
Pt. luteovirens (42)
Pt. layardi (43)

(2) Detailed classification

Genus Ptilinopus Swainson, 1825

Medium to small fruit-pigeons (wing-length 170 to 90 mm.) almost always with a
plumage pattern including well-defined patches of colour forming conspicuous
ornaments on the head and underparts, without a conspicuous yellow stripe on
the wing, and with the first primary usually more or less clearly emarginate
(falcate) at the distal end.

Subgenus Leucotreron Bonaparte, 1854 (Fig. 1).

Rather large Ptilinopids (wing-length about 160 mm.), with the head, neck, mantle
and breast concolorous or simply patterned in bright colours and sharply marked off

(e) Jambu species-group
(/) Lutovirens species-group

274 SUBDIVISIONS OF THE GENUS PTILINOPUS

from the rest of the body by a narrow pale line followed on the ventral surface by a
dark transverse bar. No tendency to ornamentation of the wings. First primary
emarginate. Greater and Lesser Sunda Isles, Australia.

(i) Pt. porphyrea (Temminck, 1823). Sumatra, Java, Bali.

(2) Pt. dohertyi (Rothschild, 1896). Sumba.

(3) Pt. cincta (Temminck, 1810). Lesser Sunda Isles except Sumba, from

Bali to Teun, Damar and Babar, and Northern Territory of Australia.

These three species are very closely allied, and if the very small overlap of Pt.
porphyrea and cincta on Bali may be disregarded, they can be grouped as the Pt.
cincta superspecies.

Subgenus Ramphiculus Bonaparte, 1854 (Fig. 1)

Medium-sized to large Ptilinopids (wing-length 150 to 170 mm.) with the most
deeply coloured patches of the head-ornamentation (excepting the chin-stripe)
lateral, sometimes meeting on the hind neck to form a ring. (A red cap occurs in
Pt. merrilli faustinoi.) No tendency to ornamentation of the wings, nor to a pale
line bounding the whole of the anterior parts as in Leucotreron. First primary
emarginate, usually very obviously. Philippines and Celebes.

(a) Occipitalis species-group

Rather highly ornamented forms (plus one with a simplified pattern, Pt. merrilli),
with a complex pattern on the head in which a chin stripe is not a well-marked
feature, with breast and belly usually of different colours separated by a dark
transverse abdominal line or band, and with spotted under tail coverts.

(i) Marchei subgroup

With a red cap and black auriculars, or only a cap, or no ornamentation on the
head. Outer web of the secondaries composed of short widely separated red barbs.

(4) Pt. marchei (Oustalet, 1880). Philippines ; Luzon and Polillo.

(5) Pt. merrilli (McGregor, 1916). Philippines ; Luzon and Polillo.

(ii) Occipitalis subgroup.

With a red or black band joining the red auriculars across the nape. One super-
species.

(6) Pt. occipitalis G. R. Gray, 1844. Philippines.

(7) Pt. fischeri Bniggemann, 1876. Celebes.

(b) Leclancheri species-group

Rather plain species with a well marked dark chin-stripe, and plain (unspotted)
brown under tail coverts. One superspecies.

(8) Pt. leclancheri Bonaparte, 1855. Philippines.

(9) Pt. subgularis Meyer and Wiglesworth, 1896. Celebes, Peling, Banggai

and Sula Mangoli.

SUBDIVISIONS OF THE GENUS PTILINOPUS 275

Subgenus Ptilinopus

Small to medium Ptilinopids (wing-length 90 to 160 mm. only in Pt. huttoni (26)
170 mm.), with strong tendencies to ornamentation of the wing coverts and scapulars
and of the underparts, either a pectoral or abdominal patch or both being almost
always present. The most deeply coloured patches on the head (excepting the chin
stripe) are dorsal and median, forming a cap. First primary often clearly emarginate,
sometimes obscurely so or merely tapering to the tip. Principally the New Guinea
region and islands to the eastwards.

(a) Ornatus species-group

Forms rather large for this subgenus (wing length 135-160 mm.), highly orna-
mented (except for Pt. tannensis (14), an isolated form with a simplified pattern).
Wing coverts spotted with pink, grey or white. There is a strong tendency to
grey or olive yellow on the neck and breast. Abdomen plain green or with only a
pale abdominal patch. Abdominal spot or bar (dividing the abdomen from the
breast) always absent. Under tail coverts spotted. First primary not or only very
slightly emarginate.

(i) Wallacii subgroup

With a red or orange cap, grey-spotted wing coverts, white chin and cheeks, and
grey breast.

(10) Pt. wallacii (G. R. Gray, 1858). Babar, Timorlaut, Kei and Aru Isles.

(11) Pt. aurantiifrons (G. R. Gray, 1858) . New Guinea, Western Papuan Islands,

and Aru Isles.

(ii) Ornatus subgroup

With an olive-yellow cap (red in one subspecies), a grey chin, and olive-yellow
breast. Wing coverts spotted with grey or bright pink.

(12) Pt. ornatus (Schlegel, 1871). New Guinea.

(13) Pt. perlatus (Temminck, 1835). New Guinea.

(iii) Tannensis subgroup

Head olive-yellow, most of the rest of the body plain green. Scapular spots
white. Related to the ornatus subgroup.

(14) Pt. tannensis (Latham, 1790). New Hebrides and Banks Islands.

(b) Purpuratus species-group

Small to medium species (wing-length 95 to 145, 170 in Pt. huttoni (26)) with a
bright red, purple or blue cap bordered behind with yellow (or with vestiges of such
a cap), and with clearly or obscurely bifid breast-feathers. Scapular and wing-
covert spots present, dark blue, pink, pale purple, or emerald green, often not clearly
marked. Abdomen ornamented with a large patch and a darker transverse bar or
spot, reduced or absent in a few forms. Under tail coverts plain red, orange or

276 SUBDIVISIONS OF THE GENUS PTILINOPUS

yellow (spotted only in Pt. superbus). First primary emarginate, almost always
very clearly.

(i) Superbus subgroup

Lower neck and upper back with a more or less extensive bright brownish red
or dark red band. Wide transverse abdominal band present. One superspecies.

(15) Pt. superbus (Temminck, 1810). Celebes and the Sulu Archipelago to the
Solomons and eastern Australia.

(16) Pt. perousii Peale, 1848. Fiji, Tonga and Samoa.

(ii) Purpuratus subgroup

With no red on the lower neck and upper back, and no distinct humeral patch
on the wing. With a narrow abdominal band transversely elongated (Pt. pulchellus
(19)) or, much more frequently, a rounded or longitudinally elongated abdominal
spot. One superspecies with two doublets (see p. 283).

(17) Pt. monacha (Temminck, 1824). North Moluccas (Halmahera, Ternate
Bat j an).
18) Pt. coronulatus (G. R. Gray, 1858). New Guinea, Japen, Salawati and
Aru Isles.

(19) Pt. pulchellus (Temminck, 1835). New Guinea and Western Papuan
Islands.

(20) Pt. regina Swainson, 1825. Eastern and northern Australia, eastern Lesser
Sunda Isles, Banda and Kei Isles.

(21) Pt. roseicapilla Lesson, 1831. Marianas.

(22) Pt. greyii Bonaparte, 1857. New Caledonia to Santa Cruz Islands and
Gower Island.

(23) Pt. richardsii (Ramsay, 1882). Ugi, Santa Anna, and Rennell Island
(Solomons) .

(24) Pt. porphyraceus (Temminck, 1821). Fiji, Tonga, Samoa, Caroline and
Palau Islands.

(25) Pt. rarotongensis Hartlaub and Finsch, 1871. Rarotonga.

(26) Pt. huttoni (Finsch, 1874). Rapa Island.

(27) Pt. purpuratus (Gmelin, 1789). Society Isles and Tuamotu or Lau Archi-
pelago.

(28) Pt. insularis (North, 1908). Henderson Island.

(29) Pt. mercieri (Des Murs and Prevost, 1849). Nukuhiva and Hivaoa (Mar-
quesas) .

(30) Pt. dupetithouarsii (Neboux, 1840). Marquesas, widespread.

Forms medium sized for this subgenus (wing-length 1 15-135 mm.). Pectoral
patch large, sharply defined and coloured white, yellow, or deep red. Abdominal
patch reduced and dark purple, or absent. Wing-covert spots present, dark blue
or grey. First primary not or only very slightly emarginate.

SUBDIVISIONS OF THE GENUS PTILINOPUS 277

(i) Rivoli subgroup

Wing spots dark blue. Abdominal patch present. Red or purple cap present,
sometimes reduced to a pair of supraloral spots. Under tail coverts yellow.

(31) Pt. solomonensis (G. R. Gray, 1870). Solomon Islands, Bismarck Archi-

pelago and some islands in Geelvink Bay.

(32) Pt. rivoli (Prevost, 1843). South Moluccas, Western Papuan Islands,

north-west New Guinea, islands in Geelvink Bay, some islands off
south-eastern New Guinea, Solomon Islands.

(ii) Viridis subgroup

Wing spots grey. Abdominal patch and red cap absent. Head shades of green
and grey (rarely white) . Size of pectoral patch varies greatly. Under tail coverts
spotted.

(33) Pt- viridis (Linnaeus, 1766). South Moluccas, Western Papuan Islands,

north-west New Guinea, islands in Geelvink Bay, some islands off
south-eastern New Guinea, Solomon Islands.

(d) Hyogastra species group

Forms medium-sized to small for this subgenus (wing-length 90-130 mm.).
Breast, neck and back plain green, unornamented. Abdomen with an orange or
violet patch (absent in Pt. melanospila (39)). Head plain green or grey, unorna-
mented or with a chin stripe and nuchal spot. First primary indistinctly emar-
ginate or merely tapering.

(i) Iozonus subgroup

Head green with greyish ill-defined chin stripe. Abdominal patch large, orange.
Grey patch on bend of wing, grey spots on wing-coverts. Under tail coverts spotted.
One superspecies.

(34) Pt. iozonus (G. R. Gray, 1858). New Guinea, Western Papuan Isles and

Aru Isles.

(35) Pt. insolitus (Schlegel, 1863). Bismarck Archipelago.

(ii) Hyogastra subgroup

Head grey (green in Pt. naina). Abdominal patch violet (absent in Pt. melano-
spila). Under tail coverts yellow, or yellow grading to red, not spotted. One
superspecies. (Possibly Pt. melanospila should be kept separated.)

(36) Pt. hyogastra (Temminck, 1824). Halmahera and Batjan (North Moluccas).

(37) Pt. granulifrons Hartert, 1898. Obi Major.

(38) Pt. naina (Temminck, 1835). New Guinea and Western Papuan Isles.

(39) Pt. melanospila (Salvadori, 1875). Philippines, Celebes, Java, Lesser

Sunda Isles (Bali to Alor) and Ceram.

278 SUBDIVISIONS OF THE GENUS PTILINOPUS

(e) Jambu species group

Head red with black chin stripe. Upper parts green, underparts white with a
pink flush on the upper breast and brown under tail coverts. First primary emar-
ginate.

(40) Pt. jambu (Gmelin, 1789). Malay Peninsula, Sumatra, Borneo and

islands between.

(/) Luteovirens species group

Rather small forms (wing-length about 115 mm.). Head more or less olive-
yellow, rest of body almost uniform green, orange or yellow, with no colour-orna-
ments. Some contour feathers lax and hairy or bifid, or long and thickened. First
primary not emarginate. One superspecies, confined to Fiji.

(41) Pt. victor (Gould, 1872). Vanua Levu, Taviuni, Kio Rambi, Ngamea,

Lauthala.

(42) Pt. luteovirens (Hombron and Jacquinot, 1841). Viti Levu and nearby

islands.

(43) Pt. layardi (Elliot, 1878). Kandavu and Ono.

NOTES ON THE CLASSIFICATION

(1) Generic and subgeneric limits

Peters (1938) divides the species considered in the present paper into three genera,
Leucotreron (i.e., the subgenera Leucotreron and Ramphiculus) , Ptilinopus (the sub-
genus Ptilinopus without the luteovirens species-group), and Chrysoena (the luteo-
virens species-group), which is divided into two subgenera. In this he differs from
Salvadori only in promoting Leucotreron from a subgenus of Ptilinopus, and removing
Pt. jambu (40) from it, in using no subgenera of Ptilinopus (Salvadori recognized
twelve), and in accepting Wetmore's subdivision of Chrysoena (1925 : 833). Amadon
(1943) has shown conclusively that Chrysoena must be ranked as a single super-
species (41-3), which requires no subdivision, of Ptilinopus. The separation of
Leucotreron is justifiable since it differs sharply from the other groups recognized
as subgenera of Ptilinopus by Salvadori, but when these are ranked only as species-
groups, there is no good reason why Leucotreron should be generically separate.
Moreover, as the classification given above will show, Leucotreron itself requires
division. A glance at the diagrams indicates that the major division of the genus
is into three, not two, groups, which have very different colour patterns and distri-
butions ; accordingly, all three are recognized here as subgenera. This arrangement
is the more satisfactory since the colour pattern of the subgenus Leucotreron, as
recognized here, is so like that seen in both Drepanoptila (as Peters recognized) and
in some of the most ornamented species of Ducula, that a complete revision of all
the fruit pigeons may possibly show that Leucotreron should be separated generically
from Ramphiculus and Ptilinopus.

SUBDIVISIONS OF THE GENUS PTILINOPUS 279

The characters by which both Peters and Salvadori define their group Leucotreron
(i.e., Leucotreron plus Ramphiculus) are the relatively long tail and the absence of a
sharply-defined red or violet cap. Peters adds that the tarsus is feathered for more
than half its length, there are no spots on the scapulars, tertials or wing coverts,
and a humeral patch is absent, and refers Pt. jambu (40) not to Leucotreron but to
Ptilinopus. In all these forms, the length of the tail tends to be associated directly
with body size, which can vary considerably within a single species-group. The
feathering of the tarsus is variable, but it is true that it is on the whole more exten-
sive in both Ramphiculus and Leucotreron than in Ptilinopus. All the other characters
are negative. Consequently, what Ramphiculus and Leucotreron share is mainly
medium to large size and the absence of the distinctive characters of Ptilinopus. On
such grounds as these it is difficult to see why Ptilinopus and Leucotreron should
not be combined instead, in opposition to Ramphiculus, an arrangement which has
never yet been proposed ; but even a cursory inspection of the positive characters
of Leucotreron and Ramphiculus shows that all these subgenera must be separated.

The only species difficult to place is Pt. jambu (40), which shows an extraordinary
mixture of characters. Because of its brown under tail coverts and black chin
stripe Salvadori associated it with Pt. leclancheri (8) and subgularis (9), its emar-
ginate first primary, lack of a cap, and rather long tail being sufficient to include it
in his Leucotreron. Peters merely remarks that by his definitions it is a Ptilinopus
(1938 : 378), but it is not easy to understand this remark, since his diagnosis of
Ptilinopus is " tail less than seven-tenths of wing, often less than six-tenths, usually
a trifle over six-tenths ; tarsus never feathered for more than three-quarters of its
length, seldom over one-half ; a sharply-defined red or violet cap ; spots on scapu-
lars, wing-coverts or tertials ; bend of wing often differently coloured from the
back," while for his Leucotreron he gives " tail more than seven-tenths length of
wing (usually more than 75 per cent.) ; tarsus feathered for more than half its
length (usually from three-quarters to completely) ; no sharply-defined red or
violet cap ; no spot on scapulars, wing coverts or tertials ; bend of wing concolour
with back (no humeral patch)."

In fact, the curious distribution of red on the head of Pt. jambu could be derived
either by extension of the red cap (and sometimes malar spots) of some Ptilinopus
or by reduction of the pattern found in either Leucotreron or Pt. marchei (4). The
chin stripe, unicolorous under tail coverts, pale underside, and lack of wing-orna-
ments are very reminiscent of Ramphiculus (the leclancheri species-group) but also,
except for the pale underside, of Pt. melanospila (39) (Ptilinopus, hyogastra species-
group). The distribution could be the result of a westward invasion from Celebes
by either of these species-groups, or indeed of an extension from Borneo or Malaya
by a geographical representative of Pt. porphyrea, producing a double invasion of
Sumatra.

Garrod (1874) states that Pt. jambu (40), Pt. perousii mariae (purpuratus species-
group (16)) and Pt. melanospila melanauchen agree in the structure of the gizzard
and differ from Treron calva. Cadow (1933) who corrects and extends Garrod's
observations, compared Pt. cincta (3), porphyrea (1) and dohertyi (2) (all Leucotreron)
and Pt. jambu (40) with species of Megaloprepia, Ducula, Treron, Columba, and

2 8o SUBDIVISIONS OF THE GENUS PTILINOPUS

Didunculus. He concluded that three main types of gizzard could be distinguished
in the fruit pigeons, and that Pt. jambu agreed in this respect with the sub-genus
Leucotreron. It seems, therefore, that the structure of the gizzard is much the same
within the genus Ptilinopus, in the subgenera Leucotreron and Ptilinopus, and in
Pt. jambu, and consequently gives no information about the position of Pt. jambu.
On the whole, the characters of Pt. jambu seem to me to suggest an association
with Pt. melanospila (39) and thence with the rest of the subgenus Ptilinopus, but
I have placed it in a species-group of this subgenus only with the greatest hesitation.
It is most remarkable that the species situated geographically at the point of con-
vergence of the ranges of the subgenera of Ptilinopus should show such a mixture of
the characters of all three.

(2) Leucotreron and Ramphiculus

In discussing the species of these two subgenera Peters (1938 : 378) was misled
by the bifid breast feathers of Pt. porphyrea (1) and the rich red on its head and
breast into considering it most closely allied to the subgenus Ptilinopus, and to Pt.
occipitalis (6) and Pt. marchei (4), the most ornamented members of Ramphiculus.
Consequently he proposed an artificial arrangement beginning with Pt. porphyrea (1),
followed by the occipitalis species-group (4-7), then the leclancheri species-group
(8-9), and finally Pt. cincta (3) and dohertyi (2) which he regarded as highly specialized.
" The more I study ductus and dohertyi," he writes, " the more apparent it becomes
that these two species are the most specialized members of the genus [his Leuco-
treron^ ; it is also evident that in spite of their superficial dissimilarity to each
other in colour, they are certainly derived from the same ancestral stock, the densely
feathered tarsi, proportion of wing to tail, modification of the inner primaries and
the pale anterior part of the body sharply defined from the dark posterior, all point to
some common ancestor." His failure to associate them with Pt. porphyrea in spite
of these very apposite remarks is in agreement with the fact that in the Checklist
(1937) he arranged the members of the purpuratus species-group in a very similar
and artificial way, bringing together the most highly ornamented forms with no
regard for the geographical evidence.

Bifid breast-feathers occur independently (as Peters points out, p. 388) in several
groups of pigeons, including (to take examples only from the fruit-pigeons) species
of the purpuratus species-group and luteovirens species-group in the subgenus
Ptilinopus, and Ducula goliath. They vary much in degree of development even
among closely related forms ; even if they do represent the retention of an ancestral
character, it is evident that the common ancestor possessing them must be a long
way back in the lineage of the pigeons. Certainly those species that possess them
to-day are not closely related. The whole pattern of ornamentation of Pt. por-
phyrea links it not with the subgenera Ramphiculus or Ptilinopus, but with Pt.
dohertyi and Pt. cincta.

Even if richness of ornamentation can be considered as one " character," it is
not a useful character in this genus since it has evidently been lost (or perhaps
gained) independently several times, as shown in the following table :

SUBDIVISIONS OF THE GENUS PTILINOPUS

281

Ornamentation in Closely Allied Species

Species-group.
Occipitalis .

Ornatus
Purpuratus .

Complex.

Pt. marchei (4)

Pt. occipitalis (6)

Pt. leclancheri (8)

Pt. ornatus (12) and

perlatus (13)

Pt. superbus (15)

Pt. regina (20)

Simple.

Pt. merrilli (5)

Pt.fischeri (7)

Pt. subgularis (9)

Pt. tannensis (14)

Pt. perousii (16)

(see Cain, 1954)

Pt. purpuratus (27)

and geographically

adjacent forms.

The luteovirens species-group (41-3) also consists entirely of a group of long-
isolated forms with extremely simple patterns derived from the subgenus Ptilinopus
(Amadon, 1943) which contains principally highly ornamented forms. Variation
between a complex and a more simple pattern has therefore occurred at least seven
times independently in the genus (and certainly involves a simplification in three
examples). But, more important, " richness of ornamentation " asa character is
clearly applicable to species which agree only in that they are richly ornamented
and differ profoundly in their patterns of ornamentation. Pt. porphyrea (1), mavchei
(4) and superbus (15), for example, are all richly ornamented, but their affinities are
not with each other but with Pt. dohertyi (2), merrilli (5) and the purpuratus species-
group (15-28) respectively ; to class them together on the basis of this very super-
ficial " character " is to ignore entirely the wealth of evidence on their real affinities
provided by their pattern of ornamentation. Colours often vary greatly in intensity
between subspecies, still more between species, but the pattern tends to remain
constant in most species-groups. There is no doubt that the pattern of Pt. porphyrea
(1) links it with Pt. dohertyi (2) and Pt. cincta (3).

The three species of Leucotreron are geographical representatives except for a
single overlap on Bali, and can be regarded as a single superspecies. The loss of
bright red pigment in the Lesser Sunda Isles form (Pt. cincta (3)) is paralleled by
the reduction of red and yellow in the subspecies of Pt. regina (20) (purpuratus
species-group) in the same area. No other species of the genus Ptilinopus is found,
in these islands except Pt. melanospila (39), which is obviously a very recent arrival
from Celebes.

The species of Ramphiculus fall naturally into two ecological groups, the marchei
subgroup (4-5) being confined to mountain forest in the Philippines (Delacour and
Mayr, 1946), while Pt. occipitalis (6) and lechlancheri (8) are lowland-forest species
both of which have colonized Celebes and produced there simplified forms with no
ornamentation on the breast and belly.

(3) Subgenus Ptilinopus

This subgenus can be readily subdivided into five species-groups, characterized
by the accentuation of different elements of the pattern. In the ornatus group

282 SUBDIVISIONS OF THE GENUS PTILINOPUS

there is considerable ornamentation of the wing-coverts and head, but very little
on the lower breast and belly. In the purpuratus group the cap, breast-patch,
abdominal band or spot, and abdominal patch are conspicuous. In the viridis
group the breast-patch is greatly accentuated, and in the hyogastra group the
abdominal patch. The luteovirens species-group (41-3), like Pt. perousii (16)
(Cain, 1954) shows a simplified pattern, as is usual in members of reduced avifaunas.

Strong sexual dimorphism is found sporadically in the subgenus Ptilinopus (but
in the other subgenera only in Pt. leclancheri (8)) and requires discussion in relation
to specific limits. It occurs in the ornatus group only in Pt. tannensis (14) (Amadon,
1943). In the purpuratus group it is present in both members of the superbus
subgroup, and in Pt. monacha (17), which is otherwise very closely related to Pt.
coronulatus (18). In the hyogastra group dimorphism is slight or absent except
in Pt. melanospila (39) in which the female has a green head. It is also seen in
Pt. jambu (40) in which the red of the head is much duller in the female and there
is some green on the breast. In the viridis group strong dimorphism is usual, but
it varies greatly in the various subspecies of Pt. viridis (33). In this species there
is a red breast-patch, which is large in Pt. v. viridis (southern Moluccas) and present
in both sexes although perhaps very slightly smaller in the females. In Pt. v.
pectoralis (western Papuan Islands and north-west New Guinea) it is very small in
the males and absent in the females. In Pt. v. salvadorii (Japen and the adjacent
part of northern New Guinea) it is rather larger and is present, though reduced, in
the females. In fact it appears that in these subspecies the patch is always smaller
in the females than in the males and when it is very reduced in the males it is neces-
sarily absent in the females. But in Pt. v. geelvinkianus (some islands in Geelvink
Bay) it is large in the males and quite absent in the females. In Pt. v. vicinas
(D'Entrecasteaux Archipelago and Trobriand Isles), Pt. v. lewisii (Lihir Islands and
most of the Solomons), and Pt. v. eugeniae (San Christobal) it is quite large in both
sexes.

Because of this variation in dimorphism, Peters divides Pt. viridis, as understood
here, into four species, Pt. viridis, eugeniae (including lewisii and vicinus), geel-
vinkiana, and pectoralis (including salvadorii). This is not necessary. All these
forms are very closely related and all are geographical representatives, with no
overlap or contiguity of ranges. Consequently, since they never meet in the wild,
it is impossible to say whether they are species or subspecies ; when there is reason-
able doubt it is much more convenient to list such forms as subspecies, so that their
close relationship is immediately obvious from their names. Sexual dimorphism
is not necessarily a specific character. It is an individual character varying within
single populations in some of the Geospizinae (Lack, 1947).

(4) The purpuratus species-group

This species-group is the largest and the most difficult to subdivide in the genus.
It has been revised recently by Ripley and Birckhead (1942) who have introduced
many notable improvements in the arrangement of the various forms. A further
consideration of it (Cain, in press) shows that certain alterations to Ripley and

SUBDIVISIONS OF THE GENUS PTILINOPUS 283

BirckheacTs subdivisions are necessary. They omit Pt. pulchellus (19) and Pt.
superbus (15) which also belong to the group (Cain, 1954), and divide the forms into
four subgroups, (i) the " old stock," the forms in Australia, New Guinea the North
Moluccas, and the Marianas (17, 18, 20, 21), (ii) Subgroup A, the forms from Raro-
tonga eastward and Pt.perousii, (16, 25-28), (hi) Subgroup B, those in the Solomons,
New Hebrides, Fiji (Pt. porphyraceus) , Carolines and Palau Islands (22-24) an d
(iv) the two Marquesan forms (29, 30) which they think should probably be placed
in Subgroup B. The characters given by them as distinctive of Subgroups A and B
are inconstant. They were unfortunately unable to see specimens of Pt. raro-
tongensis (25). An examination of specimens in the British Museum (Natural
History) shows that it is in every way intermediate in pattern as well as geographi-
cally between Pt. porphyraceus (24)) (Subgroup B) and Pt. purpuratus (27) (Subgroup
A), and that the proposed distinction between the two subgroups cannot be upheld.
On the basis of this distinction, Ripley and Birckhead suggest that the " old stock "
has given rise to two eastward expansions, one producing Subgroup A including
Pt. perousii (16), the other Subgroup B, to the west of A, which has since spread
into Fiji (Pt. porphyraceus (24)) and there overlaps with Subgroup A without inter-
breeding. Subgroup B, they suggest, has also colonized the Marquesas twice
(producing Pt. dupetithouarsii (30) and Pt. mercieri (29)), leaping over the enormous
range of Subgroup A to do so.

The abolition of the distinction between Subgroups A and B allows us to recognize
that, with the exception of Pt. perousii (16) and the Marquesan species (29, 30) all
the forms from Pt. regina (20) eastward are geographically representative and so
closely allied that there is no reason to believe them to be more than the results
of a single vast eastward expansion (Cain, in press). Pt. perousii (16) does not
belong to this superspecies, but forms with Pt. superbus (15) a distinct subgroup
within the species-group (Cain, 1954). The closest allies of the Marquesan species
are Pt. insularis (28) on Henderson Island (regarded by Ripley and Birckhead as a
subspecies of Pt. purpuratus (27)) which appears to be closely related to Pt. mercieri
(29), and Pt. purpuratus (27) which is allied to Pt. dupetithouarsii (30). There is
no doubt (Mayr, 1940) that the Marquesan species are the result of a double invasion
by the same stock. Both, consequently, are geographical representatives of their
closest allies, but it is not possible to choose one rather than the other to add to the
superspecies, leaving one outside it because of their overlap. Consequently in the
classification given above they are placed next to their closest allies and bracketed
together as a doublet. The same procedure is used for Pt. pulchellus (19) and Pt.
coronulatus (18) for the same reason. The whole of the purpuratus subgroup can
then be accurately described as one superspecies with two doublets.

The Henderson Island form, Pt. insularis (28) although a geographical representa-
tive of all the others in the subgroup, is probably rather more closely allied to a
member of the Marquesan doublet (namely Pt. mercieri (29)) than to any of the
non-overlapping forms. But its relationships are so complex that it is retained here
as a single species, since it could probably be considered almost equally well as a
subspecies of either Pt. mercieri (27) or Pt. purpuratus (27). Whatever its detailed
relationships, it is certainly a member of the superspecies.

284 SUBDIVISIONS OF THE GENUS PTILINOPUS

SUMMARY

i. Proposed classifications of the large genus Ptilinopus are unsatisfactory,
because either they are constructed on the principle that anatomical characters
are invariably more important than colour-pattern, or they are really keys. A
classification considered to be free from these defects is given.

2. Because of considerable variation in specific characters, it may be impossible
to diagnose a very natural group of species. The group is then sufficiently defined
by a description of its principal trends of variation, and a list of its contents.

3. Examples are given of the use of the species-group as a convenient informal
taxonomic rank, indicating relationships without causing nomenclatorial upheavals.

4. The use of brackets for indicating multiple invasions by closely related stocks
is exemplified.

REFERENCES

Amadon, D. 1943. Birds collected during the Whitney South Sea expedition. 52. Notes on

some non-passerine genera, 3. Amer. Mus. Novit. : 1237.
Cadow, G. 1933. Magen und Darm der Fruchttauben. /. Orn. 81 : 236-252.
Cain, A. J. 1954. Affinities of the fruit pigeon Ptilinopus perousii Peale. Ibis 96 : 104-110.
Delacour, J., & Mayr, E. 1946. Birds of the Philippines, xv + 309 pp., 69 text-figs.

New York.
Garrod, A. H. 1874. On some points in the anatomy of the Columbae. Proc. Zool. Soc.

London, 1874 : 249-259.
Hartert, E. 1896. An account of the collection of birds made by Mr. William Doherty in

the Eastern Archipelago. VII. The birds of Sumba. Novit. Zool. 3 : 576-590 and pis.

11 and 12.
Lack, D. 1947. Darwin's Finches, x + 208 pp., 8 pis., 27 text-figs. Cambridge.
Manuel, C. G. 1936a. New Philippine fruit pigeons. Philippine J. Sci. 59 : 307-310,

1 col. pi.

19366. A Review of Philippine pigeons. II. Subfamily Ptilinopodinae. Ibid., 327-336.

Mayr, E. 1940. Speciation phenomena in birds. Amer. Nat. 74 : 249-278.

1942. Systematics and the origin of species, xiv -f- 334 pp., 29 text-figs. New York.

McGregor, R. C. 1916. New or noteworthy Philippine birds 1. Philippine J. Sci. Section

D, 11 : 269-277.
Peters, J. L. 1937. Checklist of Birds of the World, 3. xiii + 311 pp. Cambridge (Mass.).
■ 1938. Generic limits of some fruit pigeons. Proc. 8th int. orn. Congr. Oxford (1934).

371-391.
Rensch, B. 1929. Die Berechtigung der Ornithologischen systematischen Prinzipien in der

Gesamtzoologie. Proc. 6th int. orn. Congr. Copenhagen (1926), 228-242.
Ripley, S. D., & Birckhead, H. 1942. Birds collected during the Whitney South Sea

Expedition. 51. On the fruit pigeons of the Ptilinopus purpuratus group. Amer. Mus.

Novit. : 1192.
Rothschild, W. 1896. [A new pigeon from Sumba Island.] Bull. Brit. Orn. Club, 5 : xlvi.
Salvadori, T. 1893. Columbae. Catalogue of the Birds in the British Museum. 21.
Wetmore, A. 1925. In Wood, C. A., and Wetmore, A. A Collection of birds from the Fiji

Islands. Ibis (12) 1 : 814-855.

PRESENTED

□

Key to Colouring
Black or Blue-Bla.

Orange, Red, Violet I
Other Colours I

LECLANCHER

PRINTED IN GREAT BRITAIN BY
ADLARD AND SON, LIMITED,
BARTHOLOMEW PRESS, DORKING

A COLLECTION OF

MESOSTIGMATID MITES

FROM ALASKA

G. OWEN EVANS

BULLETIN OF

THE BRITISH MUSEUM (NATURAL HISTORY)

ZOOLOGY Vol. 2 No. 9

LONDON : 1955

A COLLECTION OF MESOSTIGMATID
MITES FROM ALASKA

G. OWEN EVANS

"VvJ

Pp. 285-307 ; 42 Text-figures

BULLETIN OF
THE BRITISH MUSEUM (NATURAL HISTORY)
ZOOLOGY Vol. 2 No. 9

LONDON: 1955

THE BULLETIN OF THE BRITISH MUSEUM
(NATURAL HISTORY), instituted in 1949, is
issued in five series corresponding to the Departments
of the Museum, and an Historical Series.

Parts appear at irregular intervals as they become
ready. Volumes will contain about three or four
hundred pages, and will not necessarily be completed
within one calendar year.

This paper is Vol. 2, No. 9 of the Zoological series.

PRINTED BY ORDER OF THE TRUSTEES OF
THE BRITISH MUSEUM

Issued January, 1955 Price Seven Shillings

A COLLECTION OF MESOSTIGMATID MITES

FROM ALASKA

By G. OWEN EVANS

SYNOPSIS

One new genus and six new species are described ; the classification of the Epicriina is dis-
cussed and a new family Arctacaridae is erected ; the Arctic species (females only) of the
genus Arctoseius Thor are keyed.

Investigations on the acarine fauna of the Arctic prior to 1904 have been sum-
marized by Tragardh (1904). Since this date a number of papers on the systematics
and ecology of Arctic mites has appeared particularly on the Oribatei. The most
important contributions to the study of the Mesostigmatid mite fauna have been
made by Tragardh (1910), Thor (1930), Sellnick (1940) and Haarlov (1942).

The present work deals with part of a collection of mites from Alaska deposited
in the collections of the British Museum (Nat. Hist.) by the collector, N. A. Weber,
in 1950. A preliminary account of the fauna and descriptions of the habitats from
which collections were made has been published by Weber (1950). Of the seven
species of Mesostigmatid mites present in the collection six are described as new to
science including one new genus and one new family.

The type material is deposited in the British Museum (Nat. Hist.).

MESOSTIGMATA— GAMASINA

Family Laelaptidae Berl., 1892.

Haemogamasus alaskensis Ewing, 1925

1925. Haemogamasus alaskensis Ewing, H. E., Proc. biol. Soc. Washington 38 : 138-139.
I 933- Haemogamasus stevnalis Ewing, H. E., Proc. U.S. Nat. Mus. 82 : 3.

The type specimen, a single female, was collected on a Microtus sp. at Crater
Mountain, Ophir, Alaska. It is also known from a number of localities in United
States of America (Keegan, 195 1).

The collection under study contained numerous females and nymphs from Point
Barrow.

Family Phytoseiidae Berl., 1913

Genus Arctoseius Thor

1930. Arctoseius Sig. Thor, Skr. Svalbard Ishavet Oslo, 27 : 112.

1948. Tristomus Hughes, A. M., Mites associated with stored food products, H.M.S.O., London :
I 38-i39, syn. nov.

zool. 11, 9. 17

288 A COLLECTION OF MESOSTIG M ATI D MITES FROM ALASKA

The genus Arctoseius was erected by Thor (1930) for Arctoseius laterincisus Thor
(1930), a mite collected under stones on the island of Spitzbergen. The original
description and photographs of the genotype are unfortunately inadequate for its
precise identification. Haarlov (1942) published a redescription, with figures, of what
he considered to be A. laterincisus from material collected in Greenland. Thor and
Haarlov's species are undoubtedly congeneric but without examining the type
material of the former it is not possible to be certain that they are conspecific. For
the present, however, the writer accepts Haarlov's interpretation of Thor's species.

In 1948, Hughes erected a new genus, Tristomus, which is closely related to Arcto-
seius in the structure of the venter of the female, the epistome and gnathosoma.
Recently the writer remounted and examined the genotype, Tristomus butleri
Hughes, and found that the dorsal shield in both sexes has lateral incisions charac-
teristic of Arctoseius ; the species should be removed to that genus.

The most comprehensive study of the genus Arctoseius to date was made by
Willmann (1949) who added seven species to the genus ; six of which were new to
science. Willmann also proposed the division of the genus into four subgenera,
namely, Arctoseius s. str., Arctoseiulus, Arctotarseius and Arctoseiodes. The sub-
genus Arctoseiodes (type Arctoseius {Arctoseiodes) ibericus Willm. (1949) is charac-
terized by females having a ventri-anal shield and a genital shield with a pair of
setae. It is sufficiently distinct from the Arctoseius group to warrant generic status.
The subgenera Arctoseiulus (type Laelaps (Iphis) semiscissus Berl.) and Arctotarseius
(type A . (Arctotarseius) austriacus Willm.) are separated from Arctoseius s. str. by the
occurrence of a pair of jugular plates in the former and the structure of tarsus I in
the latter.

Vitzthum (1941) placed Arctoseius in the Hypoaspidinae, but as Willmann (1949)
has pointed out, the genus is more closely related to the Lasioseius-gr oup. The sub-
family Podocininae into which Willmann (loc. cit.) transferred the genus contains
as its type genus Podocinium Berl., a typical Macrochelid. Arctoseius should be
placed in the Phytoseiinae (family Phytoseiidae) .

The following is an emended definition of the genus Arctoseius :

Small free-living mites with the entire dorsal shield incised laterally. Dorsal
shield with 31 pairs of simple setae, lateral interscutal membrane with 10 pairs of
setae. Ventral surface of female with a narrow genital shield without setae and pores.
Sternal shield with two or three pairs of setae. Met asternal setae may or may not
be situated on small platelets. Anal shield well separated from the genital. Peri-
trematal plate usually fused with dorsal shield anteriorly but may be free in those
species showing a reduction in the peritreme. In the male the sterniti-genital shield
is distinct from the large ventri-anal shield. The gnathosoma is normal for the
family. The chaetotaxy of the palptrochanter, femur and genu is (2-5-6). The
epistome is bi- or trident ate. All the legs are provided with a pul villus and two
claws. Genotype, Arctoseius laterincisus Thor (1930).

The Arctic species (females only) of Arctoseius s. str. may be separated as follows :

1. Tarsus I with a conspicuous sensory organ laterally

Arctoseius laterincisus (Thor) Haarlov, 1942.
-. Tarsus I without such sensory organ laterally . . . . . . 2.

A COLLECTION OF MESOSTIGM ATID MITES FROM ALASKA

289

2. Epistome bidentate ...... Arctoseius cetratus (Sell), 1940.

-. Epistome tridentate . . . . . . . . . . 3.

3. Peritreme extending beyond the level of coxa I . . Arctoseius ornatus sp. nov.
-. Peritreme reduced, not reaching coxa I . . . . . . 4.

4. Movable digits of chelicerae about 3 times as long as the corniculi, para-anal

setae about \ length of the post-anal seta . . Arctoseius weberi sp. nov.

-. Movable digits of chelicerae about \\ times as long as the corniculi, para-anal
setae approximately equal in length to the post-anal seta

Arctoseius multidentatus sp. nov.

A. laterincisus has been recorded from Spitzbergen (Thor, 1930) and Greenland
(Haarlov, 1942). The type locality of A. cetratus (syn. Arctoseius bispinatus Weis
Fogh, 1947) is in Iceland. This species has also been recorded from Denmark and
England under A. bispinatus.

Arctoseius ornatus sp. nov.

Female. The dorsal shield is strongly ornamented and bears 31 pairs of simple
setae distributed as in Fig. 1. The lateral incisions of the shield are distinct. The
interscutal membrane surrounding the shield is provided with ten setae laterally.

Figs. 1-2. Arctoseius ornatus sp. nov., female. Fig. 1, dorsal view. Fig. 2, ventral view.
zool. 2, 9. I7 §

290

A COLLECTION OF MESOSTIGM ATID MITES FROM ALASKA

Ventrally, the heavily ornamented sternal shield has the normal three pairs of
setae and their associated pores (Fig. 2). The anterior margin of the shield in the
paratype is indentated between the first pair of sternal setae. The posterior margin
is truncated. The third pair of sternal pores normally associated with the meta-
sternal setae are situated on the posterior margin of the sternal shield. The meta-
sternal setae are not borne on platelets. The genital shield is narrow with its pos-
terior margin slightly convex. The endopodal plates in the region of coxae III and
IV are poorly developed. The genital setae and associated pores are situated lateral
to it. The post-epigynial plates are four in number. The region between the genital

Figs. 3-6. Arctoseius ornatus sp. nov., female. Fig. 3, lateral view. Fig. 4, Gnathosoma
ventral. Fig. 5, tectum. Fig. 6, chelicera.

and the anal shield is furnished with ten setae and four small platelets. The meta-
podals are elongate. The remaining four ventral setae are placed lateral to the anal
shield — two on each side. The anal shield is almost circular in contour (102 X 99 \l)
and bears the normal three setae. The para-anal setae are less than one-half the
length of the post-anal seta.

The stigmata are situated ventro-laterally in the region of coxa IV. The peri-
treme is well developed and reaches beyond coxa I (Fig. 3) . The peritrematal plate
is relatively large for the genus and is fused anteriorly with the dorsal shield. Pos-
teriorly it extends a short distance around the posterior margin of coxa IV. The
exopodal plates are fragmentary and not fused with the peritrematal.

The tritosternum is normal for the genus. The ventral surface of the gnatho-
soma is provided with four pairs of setae distributed as in Fig. 4. The ventral

A COLLECTION OF MESOSTIGM ATI D MITES FROM ALASKA 291

groove is provided with seven rows of denticles. The corniculi are short and do not
extend beyond the anterior margin of the palptrochanter. The pedipalps are normal
and the chaetotactic formula for the first three free segments is (2-5-6). The
specialized seta on the palptarsus is two-pronged. The tectum (epistome) is tri-
dentate (Fig. 5).

The chelicerae are strongly formed (Fig. 6). The movable digit is bidentate and
the fixed quadridentate with a short pilis dentilis.

All legs are normal for the genus.

Dimensions. Length 490-528^, breadth 280-285^.

Locality. Two females from Point Barrow, Alaska {Coll. A. Weber, 1950) ;
holotype (1954.3. 19.8) and paratype (1954.3. 19.9).

Arctoseius weberi sp. nov.

Female. The dorsal shield is faintly reticulated. The chaetotaxy of the shield
and the lateral interscutal membrane is normal for the genus.

Ventrally, the anterior and posterior margins of the ornamented sternal shield are
indentated (Fig. 7). The shield bears three pairs of setae and three pairs of pores ;
the third pair of sternal pores being situated in its posterior margin. The prae-
endopodal shields, although weakly sclerotized, are relatively large. The meta-
sternal setae lie posterior to the shield. The endopodal plates are weak. The flask-
shaped genital shield is faintly sculptured and covered with minute punctures. The
genital setae and pores lie off the shield. The post-epigynial and endopodal shields
are well-formed. The eight pairs of ventral setae are distributed as in the figure.
The anal shield is broader than long (80 X 98^). The para-anal setae are about one-
half the length of the post-anal setae.

The stigma is situated ventro-laterally between coxae III and IV. The peritreme
is reduced and reaches just beyond the anterior border of coxa II (Fig. 8). The
peritrematal plate is strongly formed but not fused with the dorsal shield as in the
preceding species. The exopodals are fragmentary and not fused with the peri-
trematal plate.

The tritosternum is normal for the genus and the gnathosoma is basically similar
to that of the preceding species (Fig. 9). The corniculi are approximately one-third
the length of the fixed digit of the chelicerae. The pedipalps are normal. The
specialized seta on the palptarsus has two prongs. The tectum (Fig. 10) is tri-
dentate with the median prong divided into three short processes distally.

The movable digit of the chelicera is bidentate and the fixed multidentate with a
short pilis dentilis (Fig. 11).

All legs normal for the genus.

Dimensions. Length 540 (x, breadth 286 [i.

Locality. A single female (the holotype 1954.3. 19. 10) from Point Barrow,
Alaska (Coll. A. Weber, 1950).

Arctoseius multidentatus sp. nov.

The form and chaetotaxy of the dorsal shield is similar to the preceding species.
The lateral interscutal membrane bears the normal ten pairs of setae.

292

A COLLECTION OF MESOSTIGMATID MITES FROM ALASKA

Ventrally, the sternal shield is lightly ornamented ; the ornamentation being
strongest in the median region of the shield (Fig. 12). The three pairs of sternal
setae are normal in position as are the two anterior pairs of pores. The third pair
of pores is unusual, perhaps aberrant, in that one is situated on the sternal shield
while the other is located on the interscutal membrane between the posterior-lateral

Figs. 7-1 1. Arctoseius weberi sp. nov., female. Fig. 7, ventral view. Fig. 8, lateral view.
Fig. 9, gnathosoma ventral. Fig. 10, tectum. Fig. 11, chelicera.

A COLLECTION OF MESOSTIGMATID MITES FROM ALASKA

293

margin of the dorsal shield and the metasternal seta. The region anterior to the
sternal shield is lightly sclerotized but not differentiated into prae-endopodal shields
as in the preceding species. The large genital shield is wedge-shaped, its posterior
margin being slightly convex. The genital setae and pores lie off the shield. The
post-epigynial shields are four in number. The eight pairs of ventral setae are dis-

Figs. 12-16. Arctoseius multidentatus sp. nov., female. Fig. 12, ventral view. Fig. 13,
lateral view. Fig. 14, gnathosoma ventral. Fig. 15, tectum. Fig. 16, chelicera.

294 A COLLECTION OF MESOSTIGM ATID MITES FROM ALASKA

tributed as in the figure. The metapodals are elongate. The anal shield is irregular
in contour. The anal setae are all approximately equal in length.

The stigma is situated ventro-laterally between coxae III and IV (Fig. 13). The
peritreme is reduced and extends as far as the middle of coxa II. The peritrematal
plate is well-developed being fused into the dorsal shield anteriorly and externally
around part of coxa IV posteriorly. The exopodal plates are fragmentary.

The gnathosoma has four pairs of ventral setae (Fig. 14). The corniculi are rela-
tively long and extend beyond the anterior margin of the pedipalp trochanter by
nearly half their length. The fixed digit of the chelicera is only about one and a
half times as long as the corniculi. The tectum is tridentate (Fig. 15). The movable
digit of the chelicera is bidentate, the fixed multidentate and with a short pills
dentilis (Fig. 16).

All the legs are normal for the genus.

Dimensions. Length 605 jx, breadth 275 (x.

Locality. A single female (the holotype, 1954.3. 19.7) from Point Barrow,
Alaska {Coll. A. Weber).

MESOSTIGMATA— EPICRIINA

At present there are two conflicting views on the classification of the Epicriina
depending on the relationship between the Epicriidae and Zerconidae. Vitzthum
(1941) considers these families to be closely related and includes both in the Epi-
criina, which he defines as follows :

" Gnathosoma von oben meist nicht sichtbar. Mannliche Genitaloffnung inmit-
ten des Sternale. Rumpf meist etwas flachgedriickt, so dass Riickenflache und
Bauchflache durch eine mehr oder weniger scharfe Kante geschieden sind. Mann-
liche Cheliceren ohne Spermatophorentrager. Weibliche Genitaloffnung ein Quer-
spalt vor einem Genitiventrale, das zwei oder mehr Haarepaare tragt. — Bisher ist
nur bei Triangulozercon ein dem der Gamasides gleichendes Herz nachgewiesen."

Tragardh, on the other hand, disagrees with Vitzthum and in his important work
on the classification of the Mesostigmata, based primarily on the structure of the
sterniti-genital region in the female, places the Zerconidae in a distinct division, or
cohort, the Zerconina. In Tragardh (1938) the Epicriidae are included in the
Sejidae (= Liroaspidae) a family of the Sejina (= Liroaspina) on the basis of the
females possessing a primitive genital shield. This relationship between the Epi-
criids and Liroaspids was retained by Tragardh (1946a) in a revised classification of
the Mesostigmata, but later (Tragardh, 19466) the Epicriidae were removed from the
Liroaspina because of the differences in structure of the genital region of the male.
This resulted in the Epicriina being grouped with the Zerconina since both had the
male " aperture closed by a biarticulated plate attached at the anterior margin."

Tragardh's separation of the Epicriina and Zerconina appears therefore to be
based on a difference in the structure of the sterniti-genital region of the females, which
he considered to be fundamental. Recently the writer has compared the structure
of the genital plate in a number of species of Epicrius s. lat. and Zercon. In both
genera the function of the plate is basically the same as in the Laelaptoidea, the only
difference being in the reduction of the hyaline epigynial portion of the plate which

A COLLECTION OF ME SOSTI GM ATI D MITES FROM ALASKA 295

is so well-developed in the free-living Laelaptoidea and Macrocheloidea. In Epi-
crius and Zercon the epigynial portion of the plate is reduced to a narrow chitinized
rim so that the egg is not extruded along a " chute " as in the Laelaptoidea. Further,
the males of both Epicrius and Zercon agree in a number of characters, for example,
in the position of the genital orifice, in the presence of a pair of hairs on the anterior
plate covering the genital orifice, and in the absence of a spermatophoral process on
the chelicerae. In view of the apparent close relationship between the families the
writer follows Vitzthum's concept of the Epicrhna and proposes the following classi-
fication : Division Epicriina : Mites with the epigynial portion of the genital plate
in the female reduced to a narrow chitinised rim overlapping the genital orifice.
Male genital opening situated in the sternal shield in the region of coxae II and III
and closed by two plates, the anterior of which bears a pair of hairs. Chelicerae
dentate in both sexes, but without spermatophoral process in the male. Pedipalps
with five free segments, specialized seta on palptarsus two or three pronged.

The Epicriina may be divided into the following superfamilies based on the orna-
mentation of the dorsal shield and the structure of the stigmata :

1. Peritreme and peritrematal plate absent in both sexes ; the stigmata enclosed
in a dorsal shield richly provided with small elevations forming a polygonal
network. Jugular plates well-developed in both sexes

superfam. Epicrioidea nov.

-. Peritreme and peritrematal plates present, dorsal shield without characteristic

ornamentation. Jugular plates absent in the female superfam. Zerconoidea nov.

The Epicrioidea contains one family only, the Epicriidae, represented by the genera
Epicrius (with the subgenus Diepicrius Berl.) and Berlesiana Turk. The Zerco-
noidea contains the following families :

1 . Dorsal plate divided into a notocephale and notogaster of approximately equal
size. Body triangular in outline with serrated lateral margins. All legs
with claws. Peritreme reduced in length. Female genital plate with a pair
of setae. Specialized seta on palptarsus two-pronged fam. Zerconidae Berl.

-. Dorsal plate entire. Peritreme normal. Leg I in both sexes without claws.
Strong sexual dimorphism in the structure of the dorsal shield. Female
genital plate without setae. Specialized seta on palptarsus three-pronged

fam. Arctacaridae nov.

The family Zerconidae comprises three genera, Zercon Koch (syn. Triangulo zercon
Jacot), Parazercon Tragardh (syn. Trizerconoides Jacot) and Prozecon Sellnick. The
family Arctacaridae is erected for Arctacarus rostratus gen. nov. et sp. nov. described
below.

The genus Seiodes Berl. 1887 (genotype Seiodes ursinus Berl. 1887) also belongs
to the Epicriina, but whether it is referable to either of the two superfamilies here
diagnosed is uncertain pending a re-examination of the types. The genotype species,
5. ursinus, has the following characters : Dorsal shield entire, ventri-anal shield
occupying the greater part of the region posterior to coxae IV in both sexes. All legs
with well-developed claws. Femur II in the male spurred. Male genital orifice in
sternal shield between coxae III. Chelicerae of the male without a spermatophoral
process. The species named S. hystricinus by Berlese (1892) may not be congeneric
with ursinus.

296 A COLLECTION OF MESOSTIGMATID MITES FROM ALASKA

The collection from Alaska contained two species belonging to the Epicriina and
these are described below.

Family Zerconidae

The only representative of this family is a female of the genus Zercon. To date
four species of the genus have been recorded from the Arctic, namely, Zercon curiosus
Tragardh, 1910, Zercon arctuatus Tragardh, 1931, Zercon triangularis Koch, 1836, and

Fig. 17. Zercon fenestralis sp. nov., female. Fig. 17, dorsal view.

Zercon solenites Haarlov, 1942. Z. curiosus is recorded from a number of localities in
Swedish Lappland, Z. arctuatus was first collected in an old orchard at Thorshavn in
the Faroes (Tragardh, 1931) and has recently been recorded from S. England by
Evans (1953), Z. triangularis, reported from a number of localities in Spitzbergen by
Thor (1930), is a widely distributed species in Europe. These three species are
described and keyed in Sellnick (1944). The fourth species, Z. solenites, is known
only from the type locality in Northern Greenland (Haarlov, 1942) and differs from
the other known species of Zercon in having a well-developed peritreme reaching to
coxa I.

A COLLECTION OF MESOSTIGM ATID MITES FROM ALASKA

297

The species collected in Alaska appears to be new to science on the basis of the
chaetotaxy of the dorsal shields. The nomenclature for the chaetotaxy of the noto-
gaster follows that of Sellnick (1944).

Zercon fenestralis sp. nov.

Female. The dorsum is entirely covered by two shields of approximately equal
size. The anterior shield, the notocephale, bears 15 pairs of simple dorsal setae dis-
tributed as in Fig. 17. The remainder of the chaetotaxy comprises stout coarsely
serrated setae. These number six pairs, including the verticals. The lateral margins

Figs. 18-19. Zercon fenestralis sp. nov., female. Fig. 18, sterniti-genital region. Fig. 19,

peritreme and peritrematal plate.

of the notocephalic shield are serrated. Its surface is strongly reticulated and is
provided with pores. The notogaster bears fifteen pairs of setae excluding a lateral
series of seven. All the lateral setae are simple. Setae of the J series are well-
developed. J1-J4 are short, simple, and of about equal length. The stout ser-
rated setae, J5 and J6, are more than three times the length of J1-J4. Series Z
comprises six setae of which Zi and Z2 are similar to Ji and J2 in size and form. Z3
and Z4, however, resemble J5 and J6 in structure. Z3 is shorter than Z4. Z5,
situated in close proximity to J6, is simple. Series S consists of four setae on either
side. Si is simple and the same length as Ji and Zi. Setae S2-S4 are stout and
serrated. Si is closer to Zi than S2, but S2, S3 and S4 are equidistant from each
other. Pore 3 is situated a short distance in front of seta Z4 and inside the line
connecting Z3 and Z4. The distribution of the remaining pores is shown in the

2Q8 A COLLECTION OF MESOSTIGM ATID MITES FROM ALASKA

Figs. 20-23. Arctacarus rostratus gen. et sp. nov. Fig. 20, dorsal view. Fig. 21, dorsal
seta. Fig. 22, ventral view. Fig. 23, tectum.

A COLLECTION OF MESOSTIGM ATID MITES FROM ALASKA 299

figure. The anterior half of the notogaster is strongly reticulated and the posterior
half coarsely punctured. The four posterior depressions are conspicuous.

Ventrally, the sternal shield bears three pairs of simple setae and three pairs of
pores (Fig. 18). The ornamentation is strong except for an oval area in the region
of sternal setae II. The endopodal plate in the region of coxae II is fused with the
lateral margin of the sternal shield. The metasternal setae are situated on the inter-
scutal membrane posterior to the sternal shield and between them lie a pair of weakly
sclerotized plates. The genital shield, truncated posteriorly, has a pair of genital
setae. The epigynial portion of the plate is strongly sclerotized. The head of a
club-shaped vaginal sclerite (?) protrudes from under the epigynial shield. The endo-
podal plate in the region of coxae III and IV is interrupted opposite the middle of
coxae IV. The large ventri-anal shield is of the form normally found in the genus
and bears eight pairs of simple setae (excluding the paranals and the post-anal setae) .

The stigma and its short peritreme are enclosed in a large peritrematal plate trun-
cated posteriorly and fused anteriorly with the notocephalic shield. The internal
margin of the peritrematal is fused with the exopodal (Fig. 19). The peritrematal
plate has a stout serrated seta and a large pore.

The legs and gnathosoma are normal for the genus.

Dimensions. Length 465(1, breadth 375 [x.

Locality. A single female, the holotype (1954. 3. 19. 27) from Point Barrow,
Alaska {Coll. A. Weber).

Family Arctacaridae fam. nov.

Diagnosis. Dorsal shield heavily sclerotized in both sexes completely covering
the dorsum of the female. Sterniti-genital region of female normal for the Epi-
criina. In the male the sterniti-genital plate is fragmented in the region of coxae
III-IV and separated from the ventri-anal plate. Male genital orifice closed by two
plates, the anterior of which is provided with a pair of stout setae. Peritrematal
plate fused with the dorsal shield anteriorly in the female but fused along its entire
length in the male. Coxa I, femur II and IV spurred in male. Tarsus I in both
sexes without pul villus and claws. Tarsi II-IV with pul villus and claws. Chelicerae
dentate. Specialized seta on palptarsus three-pronged.

Genus Arctacarus gen. nov.
With the characters of the family. Type : Arctacarus rostratus sp. nov.

Arctacarus rostratus sp. nov.

Female. The dorsal shield, 820-83011 x 515-528P1, is chestnut brown in colour
and does not completely cover the dorsum of the mite (Fig. 20). The surface of the
shield is finely striated and the dorsal setae are setose distally (Fig. 21). The shield
has a large pair of lyriform pores postero-lateral to the vertical setae and a number
of smaller pores distributed as in the figure. The posterior region of the dorsum is

300 A COLLECTION OF MESOSTIGMATID MITES FROM ALASKA

Figs. 24-29. Arctacarus rostratus gen. et. sp. nov. Fig. 24, lateral view. Fig. 25,
Gnathosoma ventral. Fig. 26, specialized seta of palptarsus. Fig. 27, chelicera show-
ing variation in the dentition of the movable digit. Fig. 28, Distal end of tarsus I.
Fig. 29, tarsus IV.

A COLLECTION OF MESOSTI GM ATID MITES FROM ALASKA 301

not sclerotized and bears nine pairs of dorsal setae of similar form to those on the
dorsal shield. Anterior to the posterior dorsal setae lies a small, weakly sclerotized
plate incompletely divided into two equal-sized platelets. It is interesting to note
that a pair of platelets is also present in this position in some species of Veigaia which
have a larger dorsal shield (or shields) than Arctacarus.

The tritosternum is well-developed with the lacinae long and setose. The orna-
mented sternal shield bearing three pairs of sternal setae and three pairs of pores
extends from the posterior edge of coxae I to the anterior border of coxae III (Fig.
22). It is produced into a broad sclerotized band between coxae I and II and is
fused along its lateral margin with the endopodal plates. The region between the
anterior margin of the sternal shield and the tritosternum is occupied by the frag-
mented prae-endopodal plates. The general structure of the sternal region in Arcta-
carus is not unlike that in the females of Veigaia. The metasternal setae lie on the
interscutal membrane between the sternal shield and the anterior lip of the genital
orifice. In between the metasternals lies a pair of small lightly sclerotized platelets
(cf. Zercon). The endopodal plate in the region of coxae III and IV is well-developed
and extends around the posterior border of coxae IV. The genital shield is flask-
shaped, strongly ornamented and without setae ; the genital setae and pores lying
lateral to the shield. The structure of the anterior (epigynial) portion of the shield
is similar to that in Zercon. The vaginal sclerite (?), however, is not nearly as well
developed. A further comparison with Zercon can be made in the presence of a small
" pore "- bearing plate situated posterior to coxae IV. The chaetotaxy of the region
between the genital and anal shields is shown in the figure. The anal shield is pear-
shaped and bears three setae — the para-anals and the post-anal setae. The shield is
conspicuously ornamented. The chaetotaxy of coxae I-IV, is, respectively, 2-2-2-1.

The stigma lies ventro-laterally in line with the third intercoxal space (Fig. 24).
The peritreme is well-developed and extends beyond the anterior margin of coxa II.
The peritrematal plate, fused at its distal end with the dorsal shield, is broad in the
region of the first intercoxal space but narrows markedly lateral to coxae I I-IV. It
is not fused with any part of the exopodal plate. The exopodal plate in the region
of coxa IV is well-developed but in the region of coxae II and III it is fragmented.
The interscutal membrane between the dorsal shield and the peritrematal plate has
three setae and two weakly sclerotized platelets.

Ventrally the gnathosoma bears four pairs of stout setae (Fig. 25). The external
posterior rostral seta is situated midway between the anterior rostral and internal
posterior rostral setae and external to the line connecting them. The floor of the
ventral groove is provided with ten rows of denticles. The corniculi are pointed
distally and extend a short distance beyond the middle of the palptrochanter. The
chaetotaxy of the palptrochanter, femur and genu is 2-5-6. A number of the inter-
nal setae on the palpgenu, trochanter and femur are serrated distally and the two
internal setae on the palp spatulate. The specialized seta on the palp tarsus is
three-lobed (Fig. 26). The tectum (epistome) is large and triangular in form with
five distinct projections (Fig. 23).

The dentition of the movable digit is variable. Fig. 27 shows a unidentate and
mult ident ate form ; the latter having nine small teeth. The fixed digit has two

302

A COLLECTION OF MESOSTIGM ATID MITES FROM ALASKA

strong teeth and five to six smaller ones. The pilis dentilis is of an unusual form
(see figure) .

Leg I, varying between I030(x and 1045 (jl in length, is approximately equal in
length to the body of the mite. Tarsus I is without pulvillus and claws (Fig. 28).
The majority of the dorsal setae on leg IV are stout and spiculate distally (Fig. 29).

One of the females examined contained an egg measuring 400 jx x 320 jjl.

Figs. 30-33. Arctacarus rostratus gen. et sp. nov., male. Fig. 30, dorsal view. Fig. 31,
ventral view. Fig. 32, leg I. Fig. 33, leg IV.

A COLLECTION OF MESOSTI GM ATI D MITES FROM ALASKA

303

Male. The contour of the body differs markedly from the female. The dorsal
shield covers the entire dorsum of the male. The shield is broadest at the shoulders
and tapers strongly towards the posterior end of the body. The chaetotactic pattern
and the distribution of pores is shown in Fig. 30. The dorsal setae are serrated.

The tritosternum is basically the same as in the female (Fig. 31). The sterniti-
genital region is provided with a number of plates. The prae-endopodals are elongate
and distinctly separate from the sternal shield. Sternal setae I are situated on the
heavily ornamented anterior portion of the plate and are separated from it by a
narrow strip of less strongly chitinized cuticle. Sternal setae II and III and pores I
and II are placed in the weakly ornamented portion of the sternal shield extending
from the middle of the first intercoxal space to the posterior margins of coxae II.

Figs. 34-36. Arctacarus rostratus gen. et sp. nov., male. Fig. 34, lateral view. Fig. 35,

tectum. Fig. 36, chelicera.

The shield is not produced between coxae I and II as in the female. The genital
orifice is situated in the posterior third of the shield and is closed by two plates ; the
anterior of which bears two stout setae. The shield in the region between coxae II
and III is fragmented and variable in form. The lateral margin of the shield is fused
with the endopodal plate. In the type specimen the shield is divided longitudinally
into two shields of unequal size and each shield bears two setae and the metasternal
or third sternal pore. The shields are punctured. The ventri-anal shield has four
pairs of pre-anal setae in addition to those setae associated with the anal region,
namely, the para-anals and the post-anal setae.

3<M

A COLLECTION OF MESOSTIGM ATID MITES FROM ALASKA

The peritrematal plate is fused along its entire length with the lateral margin of
the dorsal shield (Fig. 34). The stigma is situated laterally in the region of the third
intercoxal space. The peritreme is strong and extends beyond the anterior margin
of coxa II. The exopodal plate of coxa IV is well-developed and entire. The
remainder of the exopodal plate is fragmented.

Figs. 37-39. Dinychus micropunctatus sp. nov., female. Fig. 37, dorsal view. Fig. 38,
gnathosoma ventral. Fig. 39, tectum.

The gnathosome and palps are the same as in the female. The epistome (tectum)
on the other hand is considerably more strongly developed than in the female and is
beak-like in form (Fig. 35).

The fixed digit is bidentate and bears a short spine-like pilus dentilis (Fig. 36).
The movable digit, without spermatophoral process, is unidentate.

A COLLECTION OF MESOSTIGM ATID MITES FROM ALASKA

305

Leg I is long and slender as in the female and is without pul villus and claws. Coxa I
is provided with a strong spur (Fig. 31). Legs II-IV terminate in a pulvillus and
two claws. The femur of legs II and IV are spurred (Figs. 32 and 33).

Dimensions. Female : Length io5o-io8opi, breadth 5 15-528 fx. Male : Length
925-940 [x, breadth 575-59° M-

Locality. Two females and four males from Point Barrow, Alaska (Coll. A.
Weber, 1950). Holotype female (1954. 3. 19. 21), Allotype male (1954. 3. 19. 23)
and paratypes (1954 . 3 . 19 . 22, 24-26) .

MESOSTIGMATA— UROPODINA

Family Prodinychidae

Dinychus micropunctatus sp. nov.

Female. The dorsal shield is oval in contour and densely covered with minute
punctations. The chaetotaxy and distribution of pores is shown in Fig. 37. The
vertex is short with the vertical setae long and stout. The two pairs of long setae
situated posteriorly on the dorsal shield are setose. The marginal shield is of the

Figs. 40-42. Dinychus micropunctatus sp. nov. Fig. 40, sterniti-genital region of female.
Fig. 41, lateral view of female showing peritreme. Fig. 42, sterniti-genital region of male.

306 A COLLECTION OF MESOSTIG M ATID MITES FROM ALASKA

same' structure as the dorsal shield. The posterior marginal shield bears two pairs
of short, stout serrated setae and two pairs of simple setae. The latter lie on either
side of the serrated setae.

The tritosternum is short and of the form shown in Fig. 40. The sternal region is
covered with minute punctuations and a faint polygonal network. The anterior
margin of the sternal shield is straight. Sternal setae I and II are situated between
coxae II ; on setae I and II on either side are place on a sclerotised ridge (Fig. 40).
Sternal setae III and IV (according to Tragardh, 1943, seta IV is the pseudosternal)
lie lateral to the large epigynial shield which extends from the middle of coxae IV
to the posterior margin of coxae II. The ornamentation of the epigynial shield is
similar to that of the sternal shield. Four pairs of pores are present in the sterniti-
genital region. The remainder of the venter of the female is typical for the genus.

The peritreme is long and convoluted. The stigma is situated ventro-lateral of
coxa III (Fig. 41). The peritreme continues a short distance posterior to the stigma.
Anterior to the stigma the peritreme is looped in the region of coxae II. A pore is
present anterior and posterior to the loop.

The four pairs of ventral setae are distributed as in Fig. 38. The capitular setae
are short and setose distally. The external posterior rostrals are long and smooth
and the internal posterior rostrals, situated anterior to the exterior posterior ros-
trals, short and simple. The anterior rostrals are long, about equal in length to the
exterior rostrals, and setose. The corniculi are short and extend to the anterior
margin of the palpfemur. The chaetotaxy of the palptrochanter, femur and genu is
2-5-5). The setae on the trochanter are stout and setose. The tectum (epistome)
is produced into a long, narrow process, bifid distally (Fig. 39). The shaft is covered
with small spines.

All legs terminate in a pulvillus and two claws and are normal for the genus.

Male. The structure of the dorsal shield and peritreme is similar to that of the
female. The sterniti-genital region is well differentiated (Fig. 42). The genital
orifice is situated between coxae IV. The distribution of setae and pores is shown in
the figure. The complete venter of the male is finely punctated.

Dimensions. Female : Length 670-693^, breadth 370-375 \i. Male : Length
682-715 n, breadth 380-385 \i

Locality. Two females and four males from Point Barrow, Alaska (Coll. Weber.).
Holotype female (1954.3. 19. 1), Allotype male (1954.3. 19.2) and paratypes (1954
3.19.3-6).

Dinychus micropunctatus sp. nov. is closely related to Dinychus sublaevis (Tragardh)
1943, in the chaetotaxy of the posterior dorsal region of both sexes. It differs from
D. sublaevis in the ornamentation of the dorsal and ventral shields, and the structure
of the sternal region of the female.

SUMMARY

1. A small collection of Mesostigmatid mites from Alaska comprises the following
seven species :

Haemogamasus alaska Ewing.
Arctoseius ornatus sp. nov.

A COLLECTION OF MESOSTIGMATID MITES FROM ALASKA 307

Arctoseius weberi sp. nov.
Arctoseius multidentatus sp. nov.
Z er con fenestr alls sp. nov.
Arctacarus rostratus gen. et sp. nov.
Dinychus micropunctatus sp. nov.

2. The Arctic species (females only) of the genus Arctoseius Sig Thor s. str. are
keyed.

3. Tristomus Hughes 1948 is made a synonym of Arctoseius.

4. The classification of the Epicriina is discussed and a new family Arctacaridae
is erected for Arctacarus gen. nov.

A cknowledgments

I am grateful to Dr. Max Sellnick, Stockholm, for comparing specimens of Dinychus
micropunctatus with the type of Dinychus sublaevis and to Dr. H. W. Parker for his
criticisms of the manuscript.

references

Berlese, A. (1887). Acari, Myriopoda et Scorpiones . . . Fasc. 38, Tav. 9.

Evans, G. O. (1954). Some new and rare species of Acarina. Proc. zool. Soc. Lond. 123:

793-811.
Haarl0v, N. (1942). A morphologic-systematic-ecological investigation of Acarina, etc.

Medd. Grenland, 128 (1) : 1-7 1.
Hughes, A. M. (1948). Mites associated with stored food products. H.M.S.O. London : 138-

139.

Keegan, H. L. (1951). The mites of the subfamily Haemogamasinae (Acari : Laelaptidae) .

Proc. U.S. Nat. Mus. 101 : 213-218.
Sellnick, M. (1940). Die Milbenfauna Islands. Goteborgs Vetensk. Samh. Handl. (5) 6b : 1-

129.

(1944)- Zercon C. L. Koch. A cari-B latter fur Milbenkunde Konigsberg. 5 : 30-41.

Thor, Sig. (1930). Beitrage zur Kenntnis der Invertebraten Fauna von Svalbard. Skr. Sval-

bard Ishavet Oslo, 27 : 1-156.
Tragardh, I. (1904). Monographie der arktischen Acariden. Jena : 1-78.
(1910). Acariden aus dem Sarekgebirge. Naturw. Untersuch. Sarekgeb. 4 (4) : 375-586.

(1931). Terrestrial Acarina. Zool. Faroes, 2 (49) : 1-69.

(1938). Further contributions towards the comparative morphology and classification of

the Mesostigmata. Ent. Tidskr. 59 : 123-158.

> (1943). Zur Kenntnis der Prodinychidae (Acarina). Ark. Zool., 34 : 1-29.

(1946a). Outlines of a new classification of the Mesostigmata based on comparative

morphological data. Acta. Univ. Lund. N.F. (2) 42 (4) : 1-37.

■ ( 1 946ft). Contributions towards the comparative morphology of the Mesostigmata (Aca-
rina) VII. Ent. Tidskr. 67 : 88-108.

Vitzthum, H. G. (1941). Acarina. Bronns Tierreich, Bd. 5, Abt. 4, Buch 5, Lief. 5 : 641-800.

Weber, N. A. (1950). A survey of insects and related arthropods of Arctic Alaska. Trans,
ent. Soc. Amer. 76 : 147-206.

Willmann, C. (1949). Das genus Arctoseius Sig Thor, 1930 (Acari). Abh. naturw. Ver.
Bremen, 32 : 349-358. ^-— «- •*

i'TED

PRINTED IN GREAT BRITAIN BY
ADLARD AND SON, LIMITED,
BARTHOLOMEW PRESS, DORKING

W **■»** IUWT

THETDTARIID SEALS OF

THE PACIFIC COAST

OF AMERICA

JUDITH E. KING

BULLETIN OF

THE BRITISH MUSEUM (NATURAL HISTORY)

ZOOLOGY Vol. 2 No. 10

LONDON : 1954

THE OTARIID SEALS OF THE PACIFIC
COAST OF AMERICA

JUDITH E. KING

Pp. 309-337 ; Pis. io-ii ; 3 Text-figures

BULLETIN OF
THE BRITISH MUSEUM (NATURAL HISTORY)
ZOOLOGY Vol. 2 No. 10

LONDON: 1954

THE BULLETIN OF THE BRITISH MUSEUM
(NATURAL HISTORY), instituted in 1949, is
issued in five series, corresponding to the Departments
of the Museum, and an Historical series.

Parts will appear at irregular intervals as they
become ready. Volumes will contain about three or
four hundred pages, and will not necessarily be
completed within one calendar year.

This paper is Vol. 2, No. 10 of the Zoological series.

PRINTED BY ORDER OF THE TRUSTEES OF
THE BRITISH MUSEUM

Issued November, 1954 Price Eight Shillings

THE OTARIID SEALS OF THE PACIFIC
COAST OF AMERICA

By JUDITH E. KING

From Alaska to Cape Horn along the Pacific coast of the Americas several species
of Otariid seals have been recognized. These include both fur seals and hair seals
which have been exploited by commercial sealers for over two centuries ; but in
spite of this there has always been some doubt about the exact identities of some of
the animals concerned.

SPECIFIC IDENTITY OF A RCT OCEPH ALU S GA LA PAGO E N S I S

HELLER

i. Heller's type specimen

During 1898-99 the Hopkins Stanford Galapagos Expedition collected some
seal skulls from Wenman Island in the Galapagos group. The skull of an adult
male animal was described by Heller (1904) as the type of Arctocephalus galapagoensis.
This skull is now in Stanford University, California, and although Heller gave its
number as 2480, Mayer (1949) gives the correct number as 2812.

Measurements of two other skulls are given by Heller : No. 2481, male, nearly
adult ; occipital crests low ; sutures indistinct. No. 2482, female, same as 2481
in age. Mayer (1949) gives the correct numbers of these as 4442 and 4446 respec-
tively.

The specific characters of the skull are given by Heller as, " Distinguishable from
its nearest ally A . philippi of Juan Fernandez by its wider skull, both the zygomatic
and mastoid measurements being considerably greater, and by its longer snout and
mandible."

Heller gives no indication of having compared this skull with that of any other
seal except A. philippi. Photographs of the type skull of A. galapagoensis (Pis. ioa
and c ; 11 a) have been compared with skulls of A. philippi and A. australis, the

ZOOL. 2, IO. l8

312 THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA

two members of the genus Arctocephalus known to occur along the western coast of
America. As Heller has stated, the skull of A. galapagoensis is markedly different
from that of A . philippi but it differs not only in the characters he mentions but also
in having a wider and less excavated palate, deep zygomatic arches and short broad
nasals. These are just the characters in which skulls of A . australis differ so markedly
from those of A. philippi and a comparison of the photographs of the skull of A.
galapagoensis with skulls of A. australis shows a remarkable similarity (Pis. 10 and n).

It should perhaps be mentioned that Heller does not indicate whether the total
length of the animal from which the type skull was derived was taken from the tip
of the snout to the tip of the tail or to the end of the longest digit of the hind
flippers. An indication that the measurement was to the latter point is gained from
dimensions which Turner (1888, Rep. Sci. Res. H.M.S. " Challenger " Zool. 26 : 39)
gives of a specimen of an adult male A . australis. They are as follows :

From snout to tip of longest digit of pes : 5 ft. 10 J in. (1,790 mm.).

From snout to tip of tail in straight line : 4 ft. 11 in. (1,498-6 mm.).

Extreme condylo-premaxillary length of skull : 233 mm.

The greatest length of the type skull is 214 mm. and the total length of the animal
is given by Heller as 1,675 mm. (5 ft. 6 in.). Thus by comparison with Turner's
measurements, the A . galapagoensis type skull length is in proportion to body length
if the latter is taken to the end of the hind flippers and not otherwise.

Measurements and proportions of the three Galapagos skulls have been compared
with those of all the A. australis skulls in the British Museum collection (Table I).
This includes 67 skulls from the Falkland Islands, four from the Straits of Magellan,
one from Lobos Island, Uruguay, two from Messier Channel, Chile, and one from
Tierra del Fuego. Graphs of the measurements of the various components of the
skulls expressed as percentages of their condylo-basal lengths gave no indication of
allometric growth. Length ratios plotted against the condylo-basal length showed no
geographical differentiation, but width ratios showed that there was some difference
between the skulls from the Falkland Islands and those from the " mainland "
population, including the Galapagos Islands. Taking the zygomatic width as an
example and using a graphical method the zygomatic width expressed as a percentage
of the condylo-basal length in the Falkland sample has a mean value of 57*2%, a
standard deviation of 2-5, and a standard error of the mean of 0-29. The correspond-
ing figures for the mainland sample including those from the Galapagos, are 61,
3-2, and 0-96. The difference between these means is 3-8 and the standard error of

the difference of the means is 1. The ratio -^ is therefore 3-8. A difference of

crd °

this magnitude would occur twice in 10,000 trials, but only once in 10,000 in the

direction observed. This is statistically significant and indicates that the A . australis

population on the Falkland Islands is slightly different from that inhabiting the

South American coast and adjacent islands, but there are not enough skulls in the

collection to see whether the individual mainland populations are in any way distinct

from one another. The zygomatic width proportions of the Galapagos skulls are

within the standard deviation of the mainland sample and slightly above that of the

Falkland Islands sample.

THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA 313

Table I. — Measurements and proportions of skulls of A. australis
(See notes on measurements used)
FALKLAND ISLANDS
Registered No. and 1949.3. 17. 1 1949.3. 17.2 1949.3. 17.3 1949

sex, if known

I949.3-I7-I
Male

3-17-4

Greatest length .
Condylo-basal length .
Zygomatic breadth
Snout length
Snout width at canines
Snout width at level of 2nd

cheek tooth
Anterior breadth nasals
Greatest length nasals
Least interorbital width*
Mastoid breadth
Length upper tooth - r

i. — m. 6 .

Registered No. and
sex, if known

Greatest length .
Condylo-basal length .
Zygomatic breadth
Snout length
Snout width at canines
Snout width at level of 2nd

cheek tooth
Anterior breadth nasals
Greatest length nasals
Least interorbital width*
Mastoid breadth
Length upper tooth - 1

i. — m. 6 .

Registered No. and
sex, if known

Greatest length .
Condylo-basal length .
Zygomatic breadth
Snout length
Snout width at canines
Snout width at level of 2nd

cheek tooth
Anterior breadth nasals
Greatest length nasals
Least interorbital width*
Mastoid breadth
Length upper tooth - row

i. — m. 6 .

mm.
248
241
155

73
52

142

100
64-3
30'3

21-6

19-9
14. 1
14-9
I3'3

58-9

92 38-2
I949-3-I7-5

mm.
239
235
132

73
54

45
30
32

25
126

100
56-2
3i'i

22-1

19-1

12-8

13-6

io-6
53-6

88 37-4
1949.3. 17.9

„ A

mm. %
240

233 IO °

138 59-2

72 30-9

50 21-5

4i
29
38
3i
129

17-6
12-4
16-3
13-3
55'4

100
59-8

30-5
23-8

19-2

i3'4
17-2
14-6
5 6-i

88 36-8

I949-3-I7-6
Male

mm.
247
239
143

73
57

46
32
41
35
134

mm.
249
242
140
73
54

45
32
35

137

100

57'9
30-2

22-3

18-6
13-2

I4'5
n-6
56-6

92 38 o

I949-3-I7- 10

t A \

mm. %

239

234 100

139 59-4

70 29-9

48 20 • 5

4i I7'5

30 12-8

38 16-2

27 n-5

126 53-8

mm.
248
246
138
75
58

45
35
36
29
126

100
5 6-l

30-5
23-6

18-3
14-2
14-6
n-8
51-2

96 39-0
I949-3-I7-7

mm.
237
233
138

72
55

34
40

29
129

100
59-2
30-9
23-6

19-7
14-6
17-2
12-4
55'4

92 39*5
I949-3-I7-H

, A .

mm. %

239

233 IO °

128 54-9
68 29-2
51 21-9

44
30
35
30
128

18-9
12-9
15-0
12-9
54-9

mm.
243
239
146
70
54

46
29
38
29
129

89
1949.

100
61 • 1
29'3

22-6

I9-2
I2-I

15-9

I2-I

53-9

37-2
3.17.8

mm.
248
244
145
74
59

5i
32
40
29
134

92
1949.

100

59-4
30'3
24-2

20-9

131

16-4
n-9
54'9

37-7
3. 17. 12

mm.
255
247
146
76
57

49
34
38

137

100
59
30
23

82 35-2 . 88 37-6 .
Posterior to supraorbital processes.

87 37*3

95 38-5

314 THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA

FALKLAND ISLANDS— continued

Registered No. and
sex, if known

Greatest length .
Condylo-basal length .
Zygomatic breadth
Snout length
Snout width at canines
Snout width at level of 2nd

cheek tooth
Anterior breadth nasals
Greatest length nasals
Least interorbital width*
Mastoid breadth
Length upper tooth - r

i. — m. 6 .

1949.3. 17. 13 1949.3. 17. 14 1949.3. 17. 15 1949.3. 17. 17

mm.
232
225

J 34

67
46

40
28
36
30
123

100
59-6
29-8
20-4

17-8
12-4
16-0
13-3
54-7

mm.
239
233
135
70

42
29
33
30
126

100

57-9
30-0

2i-5

18-0
12-4
14-2
12-9
54-1

85 37-8

87 37-3

mm.
231
225
144
68
52

44
3i
35
29
132

100

64-0

30 -2
2 3 -I

I9-6

I3'8
15.6
12-9
58-7

39*1

mm.

235

226

124

42
28

100

54-9
30-9
21 «7

18-6
12-4

34 i5*o
25 "I

115 5o-9

88 38-9

Registered No. and
sex, if known

Greatest length .
Condylo-basal length .
Zygomatic breadth
Snout length
Snout width at canines
Snout width at level of

cheek tooth
Anterior breadth nasals
Greatest length nasals
Least interorbital width*
Mastoid breadth
Length upper tooth -

i. — m. 6 .

i949-3- I 7- 18 i949-3- I 7- I 9 i949-3- I 7-20

2nd

mm.

233
227
123

67
46

37
26

113

°/

IOO
54'2

29'5
203

16-3

15-4
13-2

49-8

mm.

228

223

128

4i
29
33
29
121

100

57'4
29-1

22*4

18-4
13-0
14-8
130
54'3

mm. %

224

219 100

124 56-6

64
45

32
114

29-2
20-5

38 17-4

14-6
52-1

1949. 3-17-21

Male

mm.
225
220

134
66

42
32
33
29
125

100
60 9
30 o
22-7

19-1

14-5
150
132
56-8

87 38-3

86 38-6

86 39-3

84 38-2

Registered No. and
sex, if known

Greatest length .
Condylo-basal length .
Zygomatic breadth
Snout length
Snout width at canines
Snout width at level of 2nd

cheek tooth .
Anterior breadth nasals
Greatest length nasals
Least interorbital width*
Mastoid breadth
Length upper tooth - row

i. — m. 6 .

I949-3-I7-22 1949.3. 17.23 1949.3. 17.24 1949. 3. 17. 25

mm.
228
222

131
69
47

42
30
36

3i
117

100
59-o
3ii

21-2

18-9

I3'5
16-2

13-9
52-7

mm.
220
216
128
64
44

36
28
36
33
115

100

59'3
29-6
20-4

16-7
12-9
16-7

I5'3
53*2

mm.
220
214
124

65
46

100
57'9
30-4
21-5

88 39-6 . 83 38-4 .
Posterior to supraorbital processes.

18-2

121

33 I5'4

117 54'7

82 3 8'3

mm.

217

211

117

37
27
37
32
108

100

55*5
28-4
19-9

17-5

12-8

17-5
15-2
51-2

38-4

THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA 315
FALKLAND ISLANDS— continued

Registered No. and

i949-3- I 7-26 1949.3. 17.27 1949.3. 17.28 1949.3. 17.29

sex, if known

mm.

Greatest length .

213

214

215

219

Condylo-basal length .

209

100

210

IOO

210

IOO

214

IOO

Zygomatic breadth

118

56-5

117

55-7

114

54*3

127

59'3

Snout length

29-2

28-1

29-0

29.4

Snout width at canines

20 -6

20-9

20-0

22*4

Snout width at level of 2nd

cheek tooth

17-2

17-6

l6'2

18-2

Anterior breadth nasals

12-9

13-3

n-4

13-6

Greatest length nasals

14-8

14-5

Least interorbital width*

14-8

13-3

12-9

i5'9

Mastoid breadth

52-6

113

53-8

106

50-5 .

115

53'7

Length upper tooth - row

i — m. 6.

38-3 ■

39'5 •

38-1 .

37'9

Registered No. and

1949.

3-I7-30

1949.

3-17-31

1949.

3.17.32

1949-3

•17.33

sex, if known

Male

mm.

Greatest length .

222

218

214

215

Condylo-basal length .

218

IOO

212

IOO

212

IOO

209

IOO

Zygomatic breadth

122

55'9

115

54-2

116

54-7

115

550

Snout length

30'3

29.7

27-8

28-2

Snout width at canines

21-6

19-8

18.4

19. 1

Snout width at level of 2nc

cheek tooth

i8-8

17-5

16-5

16.7

Anterior breadth nasals

13-3

13. 2

123

"•5

Greatest length nasals

14-2

13. 2

14-8

Least interorbital width*

133

16-5

123

13*9

Mastoid breadth

108

49'5

102

48-1

102

48-1

102

48-8

Length upper tooth - rov*

i. — m. 6 .

38-9

39-6

39-2

—

Registered No. and

1949.

3.17.34

1949.3. 17.35

1949.

3-17.36

1949.3

.17.37

sex, if known

Male

mm.

(
mm.

(

mm.

Greatest length .

213

205

197

Condylo-basal length .

209

IOO

203

IOO

191

IOO

193

IOO

Zygomatic breadth

119

56-9

108

53*2

104

54'5

107

55 4

Snout length

29-2

29-1

28-3

26-4

Snout width at canines

20 -6

18.7

19-4

i8- 7

Snout width at level of 2nc

cheek tooth

17-7

16-7

I7'3

17-1

Anterior breadth nasals

n-9

n-8

I2-0

11 -9

Greatest length nasals

15-3

I4.7

15.0

Least interorbital width*

12-9

14-8

16-2

i8- 1

Mastoid breadth

107

51-2

IOI

49-8

47-6

49-7

Length upper tooth - row
i. — m. 6 .

81 38-8 . 7 8 38-4 . 76 39.8
Posterior to supraorbital processes.

71 36-8

3 i6 THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA

FALKLAND ISLANDS— continued

Registered No. and

1949.

3.I7-38

1949.

3-17-39

1949-3

.17.40

1949.

3-I7-4I

sex, if known

Male

mm.

(

mm.

r
mm.

Greatest length .

184

190

187

180

Condylo-basal length .

i8i>

IOO

187

IOO

184

IOO

177

IOO

Zygomatic breadth

54' 1 •

102

54-5 •

IOO

54'3 •

IOI

57-i

Snout length

27-6

26-2

27-2

27-1

Snout width at canines

19-9 .

16-6 .

18-5

18-6

Snout width at level of 2nd

cheek tooth

17-1 .

14-4 .

i6- 3

16-4

Anterior breadth nasals

12-7

11 -8

12-5

12-4

Greatest length nasals

14-4

16-6

15-2

15*3

Least interorbital width*

18-8

14-9

19-0

20-9

Mastoid breadth

47-5

49*7

47-8

48-6

Length upper tooth - row

i. — m. 6 .

39*2 .

37'4 •

39'i •

39-5

Registered No. and

I949.

3-17-42

1949.

3-17-44

1949.4

.17.46

1949.

3-I7-5I

sex, if known

Male

(

mm.

Greatest length .

229

243

245

222

Condylo-basal length .

225

IOO

236

IOO

239

IOO

217

IOO

Zygomatic breadth

124

55*1

140

59-3

141

58-9

121

55'8

Snout length

28-9

30- 1

30-1

29-0

Snout width at canines

20-4

22-0

21-3

20 3

Snout width at level of 2nc

cheek tooth

16-9

18-6

18-8

i7'5

Anterior breadth nasals

12-4

13-9

14-6

12-4

Greatest length nasals

151

15-3

15-9

I5'7

Least interorbital width*

10-7

9.7

12-6

n-9

Mastoid breadth

116

51-6

126

53'4

136

56-9

109

50-2

Length upper tooth - row

i. — m. 6 .

38-7

36-4

38-i

38-7

Registered No. and

1949.

3-17-53

1949.

3-17-57

1949-3

.17.58

1949.

3-17-59

sex, if known

(

mm.

Greatest length .

192

• 175

. 223

169

Condylo-basal length .

. 188

IOO

172

IOO

220

IOO

167

IOO

Zygomatic breadth

104

55'3

102

59'3

. 118

53-6

. 96

57-5

Snout length

■ 54

28-7

- 49

28-5

■ 65

29-5

• 45

26-9

Snout width at canines

• 36

19-1

• 37

21-5

19-1

17.4

Snout width at level of 2nc

cheek tooth

18-1

18-0

- 36

16-4

. 28

16-8

Anterior breadth nasals

. 24

12-8

—

13-6

13-2

Greatest length nasals

. 28

14-9

• 27

15-7

14-5

- 25

14-9

Least interorbital width*

18-6

• 35

20 -3

I4'5

—

Mastoid breadth

50-0

. 87

50-6

112

50-9

. 84

5o-3

Length upper tooth - rov

i. — m. 6 .

• 73

38-8

38-4

40-5

39*5

Posterior to supraorbital processes.

THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA 317
FALKLAND ISLANDS— continued

Registered No. and

1949.

3.17.60

1949.

3.17.62

1949-

3-I7-63

1949-

3.17.64

sex, if known

mm.

Greatest length .

. 208

220

. 213

167

Condylo-basal length .

. 204

IOO

. 216

IOO

209

IOO

164

IOO

Zygomatic breadth

. 114

55-9

• 123

56-9

120

57-4

. 94

57-3

Snout length

• 56

27-5

■ 65

30-1

28-7

• 43

26-2

Snout width at canines

. 40

19-6

• 46

21-3

21-5

17-7

Snout width at level of 2nc

cheek tooth

16-7

i8- 1

. 40

19-1

16.5

Anterior breadth nasals

• 25

123

. 28

12-9

• 27

12-9

104

Greatest length nasals

13-2

• 36

16-7

• 33

15.8

14-6

Least interorbital width*

16-7

i5'3

• 36

17-2

• 32

19-5

Mastoid breadth

102

50-0

112

51-9

107

51-2

• 85

5i-8

Length upper tooth - rov

i. — m. 6 .

40-2

39'8

39-2

39'2

Registered No. and

I949.

3.I7-65

1949.

3.17.66

1949.

3.17.67

1949.

3.17.68

sex, if known

mm.

(
mm.

°/

r
mm.

mm.

Greatest length .

231

188

180

236

Condylo-basal length .

226

IOO

184

IOO

177

IOO

230

IOO

Zygomatic breadth

119

52-7

IOI

57'i

136

59-i

Snout length

29-2

25'5

26-6

31-3

Snout width at canines

19-0

18.5

17-5

22-6

Snout width at level of 2nc

cheek tooth .

15-9

14-7

15-8

18-3

Anterior breadth nasals

—

12.5

—

Greatest length nasals

15-5

136

—

Least interorbital width*

12-8

20 • I

16-9

12-2

Mastoid breadth

107

47'3

46-2

5i-9

119

5i-7

Length upper tooth - row

i. — m. 6 .

39'4 •

38.0 .

37-3 •

39'i

Registered No. and

1949.

3.17.69

1949.

3-17.71

1949.

3.17.72

1949.

3-17-73

sex, if known

(

■»

mm.

°/
/o

mm.

Greatest length .

221

210

211

214

Condylo-basal length .

215

IOO

206

IOO

206

IOO

213

IOO

Zygomatic breadth

114

53-o

117

56-8

115

55-8

119

55-9

Snout length

30 -2

28-6

29-1

300

Snout width at canines

22-3

19-9

19-4

21 -I

Snout width at level of 2nc

cheek tooth

18 • I .

16-9

16-0

17-4

Anterior breadth nasals

II -2 .

10-7

12-6

—

— -

Greatest length nasals

I3'5

15-0

16-0

14-6

Least interorbital width*

n-6 .

14-1 .

13-6 .

12-2

Mastoid breadth

—

103

500

107

5i'9 ■

112

52-6

Length upper tooth - row

i. — m. 6 .

38-6 .

37-4 •

37'4 •

40-4

Posterior to supraorbital processes.

ZOOL. 2, 10

318 THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA

FALKLAND ISLANDS— -continued

Registered No. and

1949.

3-I7-75

1949.

3.17.80

1949- .

$.17.81

1950.

11. 14. 1

sex, if known

-A__

Male

r '
mm.

r
mm.

(
mm.

Greatest length .

216

217

197

225

Condylo-basal length .

212

IOO

213

IOO

194

IOO

218

IOO

Zygomatic breadth

115

54'2

118

55-4

109

56-2

132

6o- 6

Snout length

29-2

28-6

27-8

30'7

Snout width at canines

18-9

20 -2

18 -6

22-5

Snout width at level of 2nc

cheek tooth

16-5

16-9

15-5

18-8

Anterior breadth nasals

—

12-2

12-9

14-2

Greatest length nasals

—

i5'5

I3'9

17-4

Least interorbital width*

17-9

13-6

13-9

n-5

Mastoid breadth

108

5o-9

107

50-2

IOI

52- 1

125

57-3

Length upper tooth - ro\*

i. — m. 6 .

38-7 ■

39'9

39'2

37'2

Registered No. and

I950.

11. 14. 2

1855. 12.26. 167

1013*

sex, if known

(

mm.

(
mm.

Greatest length .

219

239

143

Condylo-basal length .

213

IOO

235

IOO

141

IOO

Zygomatic breadth

122

57'3

145

61 -7

56-7

Snout length

30-0

30-2

26-9

Snout width at canines

20-7

24-3

20 -6

Snout width at level of 2nc

cheek tooth

• 38

17.8

20-4

19-1

Anterior breadth nasals

. 28

3i'i

12-8

Greatest length nasals

• 31

14-6

17-0

14-9

Least interorbital width*

I3-I

13-6

24-1

Mastoid breadth

—

135

57-4

5i-8

Length upper tooth - row
i. — m. 6 .

5 teeth

88 37-4

57 40'4

MESSIER CHANNEL, CHILE STRAITS OF MAGELLAN

Registered No. and
sex, if known

1950. 11. 14. 4 1950. 11. 14. 3 1879. 8. 21. 5 1880. 7. 28. 12

Male

Female

Male

mm.

°/

Greatest length .

230

201

205

226

Condylo-basal length .

225

IOO

196

IOO

201

IOO

220

IOO

Zygomatic breadth

132

58-7

116

59-2

112

55-7

135

61 -4

Snout length

29'3

30-6

27-9

30-0

Snout width at canines

24'4

18-4

17-4

22-7

Snout width at level of 2nc

cheek tooth

20-4

17-9

15-9

18-6

Anterior breadth nasals

13'

13-3

10-9

—

Greatest length nasals

15-6

15-3

14-4

—

Least interorbital width*

14-2

13-3

13-4

—

Mastoid breadth

. 128

56-9

106

54-i

IOO

49-8

124

5^-4

Length upper tooth - rov

i. — m. 6 .

38-7

40-8

39-3

38-6

* Posterior to supraorbital processes.

THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA 319

STRAITS OF MAGELLAN-

continued

LOBOS

TIERRA

ISLANDS,

DEL

URUGUAY

FUEGO

Registered No. and
sex, if known

Greatest length .
Condylo-basal length .
Zygomatic breadth
Snout length
Snout width at canines
Snout width at level of 2nd

cheek tooth
Anterior breadth nasals
Greatest length nasals
Least interorbital width*
Mastoid breadth
Length upper tooth - row

i. — m. 6 .

Registered No. and
sex. if known

Greatest length .

Condylo-basal length .

Zygomatic breadth

Snout length

Snout width at canines

Snout width at level of 2nd

cheek tooth
Anterior breadth nasals
Greatest length nasals
Least interorbital width*
Mastoid breadth
Length upper tooth - row

i. — m. 6 .

1880.7.28.7

1919.7.7.
10090

1947.7. 16.4
Female

1880.7.28. 17

mm.
226
223
146
70
53

46
31
33
34
125

65
3i
23

20
13
14
15

mm.

128

210

131

126

100

207

100

129

100

63-5

131

63'3

57-4

26-9

3° -9

27-9

20 -6

. 48

23-2

21 -7

19-8

20 -3

20 -2

I3'5

15-5

13-2

I4'3

15-5

I5'5

28-6

14-9

27-1

52-4

113

54'6

5i'9

87 39

62 49-2

82 39-6

GALAPAGOS ISLANDS

2812
Type male

2481
Female

2482
Female

mm.
214
211

134

42
27'
38
25
126

100
63-5
29-4
25-7

19-9

i i3-o

18-0

n-8

59'7

mm. %

203

200 100

124 62-0

22 II'O

28 14-0
80 40-0

86 40-8

* Posterior to supraorbital processes.

mm. %
200

197 100
120 60-9

22 II-2

27 I3'7

75 38-i

54 4 x -9

The length measurements show that the mainland skulls tend to be rather smaller
than Falkland skulls of approximately the same age. Photographs of the Galapagos
skull show that it is considerably more adult in appearance than its small size would
suggest. Skulls of similar condylo-basal length from the Falkland Islands are
obviously from young animals (PL n). They have no sagittal crest, practically
no occipital crest and the zygomatic arch has not yet attained the great depth that
it does in the old animal. Heller describes the Galapagos type skull as " old adult ;
sutures largely obsolete ; occipital and parietal crests high," and the photographs
show these characters of age, as well as the deep zygomatic arch.

3 20 THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA

The fact that there is tending to be a smaller mainland race of A . australis and a
larger Falkland Island race is interesting in the light of a recent paper by Sivertsen
(1953) on a new species Zalophus wollebaeki from the Galapagos Islands. Measure-
ments of 20 skulls of this new species were compared with those of 21 Zalophus skulls
from California. The mean condylo-basal length of the Galapagos skulls was 264 mm. ,
with a maximum of 276 mm., while the corresponding figures for the Calif ornian
skulls were 292 mm. and 323 mm. The type of Z. wollebaeki, an old male, had a
condylo-basal length of 267 mm., with a high sagittal crest and very worn teeth,
while the only Californian skull of a similar age had a condylo-basal length of
323 mm.

Thus while there can be no doubt that the fur seal on the Galapagos Islands is
A. australis, the comparative measurements given here (Table I) make it seem
probable that there is a race of smaller animals living on the islands. There is not at
the moment sufficient material in the collection to be able to distinguish the Galapagos
population from other communities living on or close to the mainland of South
America, and it is for this reason that subspecific rank cannot be given to the Gala-
pagos fur seals. The type locality for A. australis Zimmermann is the Falkland
Islands, so that if this population is considered subspecifically distinct from that on
the mainland, the name should be A. australis australis. Since the individual
populations on the mainland cannot be separated subspecifically, the first available
name for the whole is A. australis gracilis Nehring 1887, an d should the Galapagos
population eventually prove to be distinct A. australis galapagoensis is available.
As the main purpose of the present paper does not principally concern the genus
Zalophus, and pending the publication of Sivertsen's more comprehensive paper
on the Otariidae, Z. wollebaeki may be regarded as bearing the same relation to
Z. californianus that the Galapagos Ar otocephalus bears to A. australis.

2. Townsend's specimens

During 1932 and 1933 several living fur seals were captured at the Galapagos
Islands and sent to the San Diego Zoological Gardens. Three of these animals
(an adult male, an adult female and a juvenile) died, and their skins and skulls form
the subject of a paper by Townsend (1934) who referred them to Arctocephalus
galapagoensis Heller. As Heller's desciption of the type of this species, particularly
of the skull is not very detailed, and he published no photograph or drawing of the
skull, Townsend seems to have based his identification mainly on the locality.
He says that he compared his skulls with those of Arctocephalus collected " in the
Straits of Magellan and on the beaches at the Galapagos and from Guadalupe,"
but gives no details of this comparison. A few measurements of the Arctocephalus
from the Straits of Magellan are inserted in a list of measurements of the male skull
from the Galapagos, but no reference is made to these in the text.

As the animals described by Townsend are fur seals of the genus Arctocephalus
the photograph of the skull of one of them has been compared with those of A.
australis and A. philippi, the only fur seals known to occur off the coast of South
America. It is at once obvious from the general similarity, and in particular from

THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA

321

the broad snout, wide palate, short nasals and deep zygomatics that the animals
received by Townsend from the Galapagos are Southern fur seals, Ar otocephalus
australis.

Townsend gives the length of the male carcase as 4 ft. 6 in. from the tip of the nose
to the tip of the tail, and the greatest basal length of the skull as 212 mm.

Townsend's measurements of the male skull are given together with those of a
skull of A. australis (1880.7.28.7) in the British Museum collection (Table II).
It is evident from the table that there is a close similarity between the two skulls.
The two measurements that show most divergence are those involving interorbital
width. With regard to these it may be said that the British Museum collection
includes specimens of length comparable to that of Townsend's with smaller inter-
orbital width, and the variability of this part of the skull is striking.

Greatest length

Condylo-basal length

Basal length (gnathion-basion) .

Basilar length (basion-henselion)

Palatal length (gnathion-palation)

Zygomatic breadth .

Canine to last upper molar (inclusive)

Distance between upper canines (internally)

Distance between 3rd upper molars (in-
ternally) ......

Interorbital width (anterior to supraorbital
process) ......

Interorbital width (posterior to supraorbital
process) ......

Width of supraorbital processes

Greatest length nasals ....

Ant. breadth nasals .

Breadth rostrum at 2nd molar .

Mastoid breadth .

Table h.

on measurements

used)

Measurements of

male skull

given

Townsend

212

208*

100

202

97-i

198

95*2

47-6

132

63-5

32-7

y) 26

12-5

I2'9

12-9

26-5

12-7

22-1

13*9

25'5

12-3

21 -2

117

56-3

australis

1880.7

28.7

226

223

100

214

95*9

209

93 7

103

46-2

146

65-5

3i'4

II -2

34
58 est.

33
3i

125

14-8

17-9

15-2
26-0
14-8

13-9

20 -6

561

Condylo-basal length estimated as it is not given by Townsend.

SPECIFIC IDENTITY OF A RCT O CEPH ALU S TOWNSENDI MERRIAM

The other fur seal of the genus Arctocephalus from the American coast, A.philippi,
was described in 1866 by Peters from the skull of an adult (but not old) male animal
collected on Juan Fernandez in December, 1864, by Dr. Philippi. Full size drawings

322 THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA

JO cms.

Fig. i. Type skull of Arctocephalus philippi from Peters, Monatsb. Akad.

Berlin, 1866, pi. 11.

THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA 323

of the skull were given (Fig. 1) from which measurements have been taken for
comparison. The most important skull characters for distinguishing the species are
given by Peters as follows :

1. The palate is deep in front and shallow behind, and is twice as broad
between the last molars as between the canines.

2. The posterior ends of the maxillae are prolonged into small hooks. 1

3. The maxillary root of the zygomatic arch is very broad antero-posteriorly.

4. The tympanic bullae are flat.

5. The mastoid process is long, but does not jut out ventrally.

6. The zygomatic arch is narrow.

7. The lower jaw has no pronounced angle ; the coronoids are rounded and
project backwards.

8. The peculiar pointed form and the lack or weak development of accessory
cusps on the molar teeth, which are separated from each other by relatively
large spaces.

All these characters are visible both in Peters' drawings of the type skull and in
a skull of this species from Juan Fernandez in the British Museum collection (Reg.
No. 1883. n. 8. 1.).

A.philippi, the type locality of which is Juan Fernandez, is a fur seal, as is evident
from Peters' description of the long overfur and the thick underfur. The only other
animal with which it may be confused is the Southern fur seal, A. australis, which
ranges from Uruquay, round the Straits of Magellan and along the west coast of
South America. Externally A. philippi may be distinguished from A. australis
chiefly by its long tapering snout which is very unlike the short, rather upturned
snout of A . australis. When seen side by side skulls of the two species show many
differences. The skull of A. philippi (Fig. 1) is more finely built than that of A.
australis (Fig. 2). In general shape it is long and slender, as are also the nasal
bones ; and the zygomatic arches are narrow in lateral view. The palate is also
narrow and is very concave between molars 1-3. The skull of A. australis is much
more robust, the nasals are short and broad, the palate is not particularly narrow or
concave and the zygomatic arches are strong and deep. These differences can
easily be seen in the drawings of the two skulls (Figs. 1 and 2) .

In 1870 the Museum at Santiago was presented with a male, a female and a young
seal from the island of Masafuera. These were considered to be a new species and
formed the subject of a paper in 1871 in which they were called Otaria argentata
(Philippi in Peters, 1871). Philippi gives a drawing of the skull of this new species
and a list of characters in which it differs from A. philippi. The skin of 0. argentata
is said to differ chiefly by its lighter colour and by the guard hairs of the neck being
shorter than those of A . philippi. The measurements given of the stuffed male show
it to be a young animal as it is only 3 ft. 9 in. from nose to tail, and Philippi himself
says that the skull is not fully grown as the crests are not developed. The length
of the neck hairs and many of the characters of the skull used by Philippi to distin-
guish the species may be explained by the youth of the specimen, and the drawing

1 Not confined to this species

3 2 4 THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA

Fig. 2. Orthographic projection of skull of Arctocephalus australis, 1880.7.28.7.

THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA

325

of the skull gives proof that the animal is another specimen of A . philippi. Peters,
in his discussion of Philippi's letter, says that if it had not been for Philippi's insistence
that 0. argentata and A. philippi were different, he " would have thought twice
about describing them as two different species."

In 1892 an expedition sailed from San Diego to Guadalupe for the purpose of
identifying the fur seals present on the island. Seven seals were seen and one was
shot, but sank. Four skulls were picked up on the beach and one of these forms the
type of a new species Arctocephalus townsendi which was described by Merriam in
1897. The skull is that of an adult male No. 83617 U. S. National Museum. Merriam
compares this type " with skulls of Arctocephalus (australis or philippi) from the
Galapagos Islands," and from the characters that he gives it is evident that he com-
pared his skull with A . australis and had not seen a skull of A . philippi.

Fig. 3. Type skull of Arctocephalus townsendi, U.S. Mus. No. 83,617. Taken
from Allen, 1905, Patag. Exp. Ill ; pi. xviii.

Merriam lists the most important characters for identifying the skull of A . townsendi
as follows :

1. The exceedingly narrow and excavated palate.

2. Flat tympanic bullae.

3. Short and thick ascending arm of premaxilla.

4. Broadly expanded zygomatic root of maxilla, forming a floor under the
anterior half of the orbit.

5. The 5th molar mainly posterior to the plane of the anterior root of the
zygoma.

326 THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA

When these characters are compared with those given by Peters for the type
skull of A. philippi it will be seen that Merriam's Nos. i, 2 and 4 — important diag-
nostic features — correspond exactly with Nos. 1, 4 and 3 respectively of Peters'
description. Merriam's character No. 3 is not visible in the ventral view of the skull
which he gives. His fifth character, while it is good for the specimen of A. philippi
in the British Museum collection, is not distinctive.

When a further comparison is made between the measurements and proportions
of the type skull of A. philippi, the skull of this species in the British Museum, and
the type skull of A. townsendi (Table III) it will be seen that these three skulls are
so similar that it is probable that they all belong to the same species of Ar otocephalus.
Comparison of the drawings (Figs. 1 and 3) of these skulls also confirms this view,
and it is thus possible to add Arctocephalus townsendi Merriam to the synonymy of
Ar otocephalus philippi Peters.

Allen (1905) described skulls from the Galapagos which he called A. philippi,
and combined a very " free translation " of the characters given by Peters with some
which are obviously only from the skulls he was describing. His figured skulls show the
typical high sagittal crest of the male Zalophus and it is clear that he has mistakenly
identified these skulls as A. philippi. Osgood (1943) says that Remington Kellogg
and G. M. Miller examined Allen's skulls from the Galapagos and found that they
were undoubtedly Zalophus calif ornianus.

Table III.

(See notes on measurements used)

A. townsendi

philippi

A . philippi

type
mm.

type

1883

11. 8. 1

mm.

Greatest length ....

256

236

• 255

Condylo-basal length

*252

100

233

IOO

. 248

IOO

Zygomatic breadth ....

151

59-9

129

55'4

139

56-0

Snout length . . . . .

t6 9

27-4

27-0

29-4

Snout width at canines .

t53

21 -o

19-3

i8-i

Canine to last upper molar (inclusive)

34'9

175

32-2

36-7

Distance between 3rd upper molars

(internally) .....

22-5

8-9

9-0

7-7

Antero-posterior width of zygomatic

root of the maxilla

8-3

8-6

8.9

Interorbital width (posterior to supra-

orbital process) ....

t3<>

n-9

12-9

12-5

Mastoid breadth ....

ti33

52-8

115

49-4

125

5o-4

Greatest length nasals

— ■

—

17-6

18 • 1

Anterior breadth nasals .

—

II -2

IO-I

* Estimated.

f Measurements taken from i/i drawing of type skull in Allen (1905). Other measurements as given
by Merriam (1897).

% Peters' type has only 5 upper molars. Measurements of skull taken from 1 /i drawing in Peters
(1866).

THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA 327

DISTRIBUTION AND PARTIAL RECORDS OF A. PHILIPPI
ON THE PACIFIC COAST OF AMERICA

The records of the occurrence of A. philippi which are listed below are based
partly on the identification of actual specimens, and partly on records of " fur seals "
from the Calif ornian Islands which are, according to Starks (1922) A. townsendi =
A. philippi.

1. Juan Fernandez

(a) 1864. Type specimen killed (Peters, 1866).

(b) — " Some years ago," possibly in 1864, a few dozen skins taken by

Philippi (Philippi in Peters, 1871).

Chilean Government.

(d) — Skull in Santiago Museum. No locality given (Philippi in Peters,

1871).

2. Masafuera

(a) 1870. 3 seals collected. One of these described as type of 0. argentata
(Philippi in Peters, 1871).

3. Chonos Islands

(a) 1871. A skin taken by Philippi's assistant (Philippi in Peters, 1871).

4. San Benito Islands

(a) 1806. 8,338 skins taken by the " Port au Prince " (Townsend, 1924, 1931).

A few hundred skins taken from Guadalupe and Santa Barbara

(Starks, 1922).

Fur seals tightly packed (Merriam 1897).

2,000 killed, but commercially extinct (Merriam, 1897).

Only a few seen (Merriam, 1897).

7 seals seen and 4 skulls obtained from which type of A. townsendi

described (Merriam, 1897).

None seen (Anthony, 1925).

2 year old female sent to San Diego Zoo (Starks, 1922).

2 males caught and sent to San Diego Zoo, and 60 seals estimated to

be on the island (Townsend, 1931).

Some fur seals seen in May (Townsend, 1930).

None seen (Bartholomew and Hubbs, 1952).

5. Guadalupe

(a)

1878.

(b) 1880.

(d) 1891.

(e) 1892 .

(/) 1922.
(g) 1922.
(h) 1928.

1929 .
1950.

328 THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA

6. Santa Barbara Islands

(a) 1878. A few hundred skins taken from Guadalupe and Santa Barbara

(Starks, 1922).

(b) 1890. 5 seals taken off San Miguel (Starks, 1922).

(d) 1949. Single adult male seen (Bartholomew and Hubbs, 1952).

7. Ventura, California

(a) Eighteenth to nineteenth centuries. Abundant remains found in shell mound
(Lyon, 1937).

8. Farallone Islands

(a) 1810-1812. 73,402 skins taken by Americans (Starks, 1922).

(b) 1812-1840. Russians took 1,200-1,500 skins annually (Starks, 1922).

(d) 1824. Seals profitable again, over 1,000 skins taken, but decreased to 54

skins in 2 years (Starks, 1922).

The first record of a relatively recent occurrence of the fur seal is the finding of
skull remains from the eighteenth to nineteenth centuries in a shell mound at Point
Mugu, Ventura County, California, 1,557 identifiable bones of this seal were found
and it was by far the most abundant animal in the collection. No more records
are available until 1806 when over 8,000 skins were taken from the San Benito
Islands. The seal was abundant on all the islands off the Calif ornian coast and was
taken for its fur in great numbers until about 1883 when it was considered extinct
commercially. From that time until the present day there have been only a few
sporadic noted occurrences, the last being an adult male seen on San Nicolas Island in
1949.

As there are so few records of A. philippi being seen on Juan Fernandez and
Masafuera it seems probable that the normal habitat of this seal is the islands off
the Calif ornian coast. When it is known to have been so plentiful in the northern
part of its range it is noteworthy that there have been so few records of its occurrence
on Juan Fernandez, although this may be attributed partly to the remoteness of this
island compared with those off the Calif ornian coast. It is probable that Juan
Fernandez represents the most southerly extent of its distribution and that there has
never at any time been a great number of A. philippi on the island. The presence of
a small number of A . philippi would be masked by the large numbers of A . australis
which undoubtedly occurred on the island, and the differences would not be noticed by
the sealers who would class them all as " fur seals." It was only scientifically in-
terested people like Dr. Philippi who collected the type which was named after him,
and the chemist Fr. Leyboldt who collected the type of 0. argentata who noticed the
difference between the two species.

The present state of A. philippi is unknown. It may be extinct, but since an
adult male has been seen as recently as 1949 it seems possible that there may still be
a small colony left.

THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA 329

Map to show seal-inhabited islands off the Pacific American coast.

1. Chonos Islands.

2. Juan Fernandez.

3. Masafuera.

4. St. Felix, St. Ambrose, St. Mary's Islands.

5. Galapagos Islands.

6. San Benito Islands.

7. Guadalupe.

8. Santa Barbara Islands.

San Nicolas.
Santa Cruz.
Santa Rosa.
San Miguel.

9. Ventura.

10. Farallone Islands.

33o THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA

PARTIAL RECORDS OF " FUR SEALS" OFF THE COAST OF

SOUTH AMERICA

1. Masafuera

(a) 1563. Island discovered. Swarming with fur seals (Allen, 1899).

(b) 1792. Capt. Stewart of the " Eliza " took 38,000 skins to Canton and sold

them for 16,000 dollars (Allen, 1899).

lion taken to Canton in 7 years. People from 14 ships on island at
the same time, all killing seals (Allen, 1899).

(d) 1798. Capt. Fanning of the "Betsy" took 100,000 skins to Canton and

estimated that when he left there were still 500,000-700,000 seals
on the island (Allen, 1899).

(e) 1807. Capt. Morrell — business scarcely worth following (Allen, 1899).

(/) 1824 . Capt. Morrell — island almost entirely abandoned by seals (Allen, 1899).
(g) 1891. Capt. Gaffney saw 300-400 and took 19 (Allen 1899).

2. Juan Fernandez

(a) 1683. Dampier — seals thick about the island (Allen, 1899).

(b) 1687-90. Capt. Davies of the " Bachelors Delight" left some men on the

island to cure seal-skins (Albes, 1914).

(d) 1800. Capt. Delano — human population of 3,000 on island, so no seals left

(Allen, 1899).

(e) 1891. Capt. Gaffney saw a few fur seals in December (Allen, 1899).
(/) 1931. Nybelin got a parasite from an A. australis (Nybelin, 1931).

3. St. Felix, St. Ambrose, St. Mary's Islands

(a) 1792. 13,000 skins taken during August and September by American ship

" Jefferson " (Howay, 1930).

(b) 1801 . Islands were visited by sealers and there must have been large numbers

of seals (Allen, 1899).

(d) 1891. No seals worth mentioning (Allen, 1899).

4. Galapagos Islands

( a ) I 535- Seals mentioned by discoverer of Galapagos (Baur, 1897).

(b) 1800. Large numbers of fur and hair seals (Allen, 1899).

(d) 1825. Capt. Morrell took a few fur seals from Albemarle (Allen, 1899).

(e) 1872 . Capt. Reed took 3,000 fur seals and about as many more during three

subsequent voyages between 1872 and 1880 (Allen, 1899).

THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA 331

(/) 1872. Hassler Expedition. Specimens of A. australis collected (Allen, 1880).

(g) 1885. Capt. Gaffney took 1,000 fur seals (Allen, 1899).

(h) 1898-99 . Capt. Noyes — seals not very numerous, took 224 skins (Allen, 1899) .

Heller collected type of A. galapagoensis on this voyage (Heller, 1904).
(i) 1923 . Beebe saw only two fur seals (Beebe, 1924)
(j) 1931. Sea-lions and seals numerous and quite tame (Korwin, 1931).
(k) 1932-33. Capt. Hancock caught several fur seals which were presented to

San Diego Zoo (Townsend, 1934).

As even the incomplete list above shows, there is quite a number of records of
sealers taking fur seals from the islands off the South American coast from the
Galapagos Islands southwards. The story of these fur seals is the usual one of
uncontrolled exploitation. Millions of skins were taken until the middle of the
nineteenth century when the numbers dropped rapidly and for about the last 40
years not only are there no records of any commercial sealing, but there are very
few referring to seals at all, and even these indicate that there cannot be many fur
seals on the islands.

It appears from the available records that the sealing expeditions worked north-
wards along the west coast of South America, as seals at Masafuera and Juan Fer-
nandez were abundant until about 1798, at St. Mary's until 1816, while there were
still considerable numbers at the Galapagos as late as 1885.

There is still some doubt as to the identity of these " fur seals/' Townsend
(in Allen, 1899) refers to the Galapagos seal as A. philippi, but does not give any
reasons for his identification. The only other records where specific names are used
are those of Allen (1880), Heller (1904), Nybelin (1931) and Townsend (1934) for
seals on the Galapagos and Juan Fernandez and in all four instances A. australis
is the animal named. Both A . australis and A . philippi have been reliably recorded
from Juan Fernandez, and A. australis from the Galapagos, so it is not unlikely,
though not proven, that A. philippi also occurs on the Galapagos. The bulk of
the records of " fur seals " from these South American islands probably refer to
A. australis although the scarcity of records makes it impossible to be certain.
For the same reason it is not known what the present status of this seal is along the
western South American coast.

Other Otariids frequent the shores and islands of the Pacific coast of America.
Of these the best known is the Calif ornian sea-lion Zalophus calif ornianus. The
most southerly point from which it has been recorded is the Galapagos Islands.
Beebe (1924) went to these islands in 1923 and makes several references to the sea-
lions that he saw there. He identifies them as " Southern sea-lion, Otaria jubata,"
but the photographs he gives show the short pointed nose of Zalophus and not the
heavy upturned snout of 0. byronia. Wollebaek (1927) went to these islands in
1925, and although he does not mention the seals by name he gives a photograph
which is undoubtedly of a Zalophus. Skulls of Zalophus from the Galapagos which
are in the U.S. National Museum are figured by Allen (1905) although he wrongly
calls them A. philippi. Zalophus is well known on the islands off the Calif ornian
coast and probably extends northwards as far as the bay of San Francisco.

332 THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA

Steller's sea-lion, Eumetopias jubatus, is known to extend its range as far south
as San Nicolas Island, although there are no records of it breeding any further south
than Santa Rosa Island, 42 miles north of San Nicolas.

The Northern fur seal Callorhinus ursinus, although it breeds mainly on the Pribilof
Islands, spends the winter and early spring at sea and may migrate southwards to
the latitude of California, although it rarely comes ashore there.

The distribution of the Southern sea-lion, Otaria byronia, corresponds with that
of A. australis. Both Heller (1904) and Allen (1880) note its presence on the Gala-
pagos Islands and it is known on Juan Fernandez, round Cape Horn and along the
eastern coast of South America as far as Lobos Island, Uruguay.

SUMMARY OF THE MAIN CHARACTERS AND DISTRIBUTION OF
THE OTARIIDS OF THE PACIFIC COAST OF AMERICA

Otaria byronia Blainville, 1820. Southern sea-lion.

Size. Length of adult male from nose to end of tail 6-7 ft., and of female 5-6 ft.
Weight of adult male over 1,500 lb.

Description. Males and females similar in colour, back usually dark brown,
mane of male and neck of female, and the belly dark yellow. Face dark and hair
of flippers reddish. Bare parts of skin black. Whiskers long, reaching 12 in.
in adult male, about 30 on each side, straw coloured and forming a pale moustache.
Variations in colour frequent, and many paler animals found. Newly-born pups
practically black, soon fading to chocolate, and after first moult at a few months old,
a dark grey.

Male with extremely thick and heavy neck thrown into folds, this together with
the lighter and thicker hair on the neck gives the appearance of a mane. Female
slighter in build. Head short and muzzle deep and upturned.

Distribution. From Lobos Island, Uruquay to Straits of Magellan, Falkland
Islands, Galapagos Islands.

Skull. Condylo-basal length up to 358 mm. in adult males, and 267 mm. in adult

females. Easily distinguishable from skulls of all other seals by the greatly elongated

palate extending backwards as far as the pterygoids, and becoming progressively more

concave posteriorly. Adult male skull with well-developed sagittal and parietal

crests, and various processes on the parietal for the attachment of muscles.

^ .316

Dentition. 1.- c- m.-.

2 1 5

Eumetopias jubatus (Schreber 1776). Steller's sea-lion.

Size. Length of adult male from nose to end of tail 11-13 ft. and of adult female
8-9 ft. Weight of adult male 1,000-1,200 lb., of adult female 400-500 lb.

Description. Both males and females a light reddish brown ; slightly darker on
the belly. Colour varying with age and season and coat lighter when just moulted.
Naked parts of skin black. Whiskers long and slender, the longest about 20 in.,
white or brownish white. Adult males with mane on neck. The largest of the eared

THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA 333

seals. Has a rather bear-like head with a short straight nose, not upturned as in
0. byronia. Muzzle and neck very heavy.

Distribution. Shores of the north Pacific from Behring Straits southward to
Santa Barbara Islands, California, and Japan.

Skull. Condylo-basal length up to 380 mm. in adult males. Skull distinguished
by the anteriorly placed quadrate supraorbital processes and, in all except young
skulls, a large gap between the upper 4th and 5th molars. Palate with hinder end
contracted and truncate.

^ .315

Dentition, i.- c- m.-.

215

Ar otocephalus philippi (Peters 1866). " Juan Fernandez " fur seal.

Size. Length of adult male from nose to end of tail ca. 5 ft. Adult female
probably slightly smaller.

Description. Above blackish grey, more yellowish grey on head and neck ;
brownish black below, proximal part of limbs, lips and chin rusty brown. Overhair
rusty brown with black tips, thick underfur rust red. Whiskers in six rows, some
black, some white, and some white with black tips. Animal distinguished by slender
tapering snout.

Distribution. Juan Fernandez, Guadalupe, Santa Barbara Islands, Farallone
Islands.

Skull : i. Skull narrow, the mastoid breadth being 48-50% of condylo-basal
length.

2. Palate narrow and very deep in front.

3. Maxillary root of zygoma very wide.

4. Flat tympanic bullae.

5. Narrow zygomatic arch.

6. Snout narrow.

7. Posterior prolongations of palatines not thickened.

8. Teeth usually without accessory cusps and widely spaced.
Measurements of few authentic specimens known but condylo-basal length of adult

male skull ca. 256 mm.

Dentition, i.- c- m.-. (Peters' type skull has only m.- but number of
molar teeth of Arctocephalus rather variable).

Arctocephalus australis (Zimmermann 1783). Southern fur seal.

Size. Length of adult male from nose to end of tail 5 ft. 6 in., and of adult female
4 ft.

Description. Overhair black, tipped with grey except on belly where blackish
brown. Hairs 1-2 in. long on back of neck but shorter over rest of body. Underfur
reddish brown. Dorsal surface of manus and pes covered up to nails with black
hairs. Tip of nose and naked skin of limbs black. Whiskers white, though some
of the smaller are greyish black, about 20 on each side. Very few photographs

334 THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA

available, but appears to be rather like Callorhinus in the external shape of the head,
except that the nose is rather longer and slightly upturned.

Distribution. Shores and islands of South America from Lobos Islands, Uru-
guay, to the Galapagos Islands, Straits of Magellan, Falkland Islands, S. Georgia.

Skull. Condylo-basal length up to about 250 mm. in adult male. Skull squarish

in general shape, heavily built, with thick zygomatic arches and short broad nasals.

-^ .316

Dentition. 1- c- m.-.
2 1 5

Callorhinus ur sinus (l.). Northern fur seal.

Size. Length of adult male from nose to the end of tail 7-8 ft. and of adult
female 4ft. Weight of adult male 500-600 lb., of adult female 80-100 lb.

Description. Males nearly black on the back and brownish ventrally, neck and
shoulders greyish. Limbs reddish brown. Naked areas of skin black. Females
lighter than males, being grey dorsally, but otherwise like males. Whiskers white,
or with brownish tips. Black in young animals. Distinguished externally by the
high forehead and extremely short pointed nose.

Distribution. From Pribilof Islands south to California and Japan.

Skull. Condylo-basal length up to ca. 250 mm. in adult males. Interorbital

region long, facial region very short and high, descending abruptly. Dentition weak.

-^ .316

Dentition. 1.- c- m.-.

2 1 5

Zahphus californianus (Lesson 1828). Calif ornian sea-lion.

Size. Length of adult male from nose to end of tail 8 ft., and of adult female 6 ft.
Weight of adult male 500-600 lb.

Description. Not so heavily built as Otaria or Eumetopias with not such a thick
neck in adult males.

Distribution. Galapagos (but see Sivertsen's (1953) account of Z. wollebaeki) ,
shores and islands of California north to Farallone Islands.

Skull. Condylo-basal length up to 320 mm. in adult males. Skull of adult

males easily distinguishable by the very high sagittal crest. Skulls of females and

young males slender with elongated nasals and facial regions.

■^ .316

Dentition. 1.- c- m.-.

215

key for distinguishing skulls of pacific american

otariidae

1. Skull with supraorbital processes and alisphenoid canal . .... Otariidae

2. Posterior end of floor of palate very concave and extending backwards as far as

the pterygoids O. byronia

ia. Posterior end of floor of palate not extending backwards as far as the pterygoids 3.

3. Supraorbital processes quadrate. Molars 5/5 with large gap between upper

m. 4 and m. 5 in adult skulls . . . . . . . E. jubatus

THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA 335

3a. Supraorbital processes triangular. Molars 6/5 with no large gap between

upper m. 4 and m. 5 in adult skulls ....... 4.

4. Nasals long and slender, combined widths at anterior ends ca. 40-50% of length 5.

4«. Nasals short and broad, combined widths at anterior ends ca. 80-90% of length 6.

5. Palate very narrow, the internal distance between m. 1-3 being the same.
Floor of palate very concave between these molars. Tympanic bullae
smooth and rounded. No very high sagittal crest in old males . A . philippi
5a. Palate wider, the distance between the molars increasing gradually.
Floor of palate only very slightly concave. Surface of tympanic bullae
irregular. Very high sagittal crest in old males . . Z. californianus

6. Interorbital region long in adult animals, ca. 20% of condylo-basal length.

Snout, at level of m. 2, practically as high as it is long . . . C. ur sinus

6a. Interorbital region less than 20% of condylo-basal length in adult animals.

Snout much longer than it is high . . . . . . A . australis

la. Skull without supraorbital processes or alisphenoid canal (not considered in this key)

Phocidae
Note : The above characters apply mainly to adult skulls.

SUMMARY

After comparison with skulls of fur seals known to occur along the Pacific coast
of America, the type skull of Arctocephalus galapagoensis, which was collected on
Wenman Island in the Galapagos group in 1898-99, is shown to be similar to skulls
of A. australis.

Skulls of fur seals brought back from the Galapagos Islands in 1932-33 are also
identified as A. australis.

A statistical treatment of proportions of skull measurements of A. australis from
the Falkland Islands on the one hand, and the mainland of South America and the
Galapagos Islands on the other, makes it seem probable that the Falkland Islands
animals belong to a larger race. The shortage of specimens makes it impossible at
the moment to distinguish the Galapagos skulls from those of other mainland popu-
lations.

It is shown by a comparison of skull drawings, measurements and proportions
that the Guadalupe fur seal A. townsendi is the same as the Juan Fernandez fur
seal A. philippi. A chronological list is given of the recorded occurrences of this
seal and it is suggested that its normal habitat is the islands off the coast of Cali-
fornia. The few records from Juan Fernandez represent animals taken at the most
southerly point of the range, but its abundance there was probably masked by the
greater numbers of A . australis.

A list is given of records of un-named " fur seals " from the Galapagos Islands
southwards. It is probable that most of these records refer to A. australis.

A brief account is given of the distribution of Z. californianus, E.jubatus, C. ursinus
and 0. byronia, the other Otariids which occur along the coast.

A summary is given of all the Otariids occurring along the Pacific coast of America,
and also a key for distinguishing their skulls.

Acknowledgments.

I wish to express my indebtedness to Professor G. S. Myers and Mr. J. M. Savage
of Stanford University, California, for the assistance they have given me by supplying

3$6 THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA

information and measurements relating to the type skull of A. galapagoensis, and
also to Dr. Antenor L. de Carvalho, Museu Nacional, Rio de Janeiro, who took the
photographs of this skull reproduced in the text, while working in Stanford University.

NOTES ON THE MEASUREMENTS USED

i. Greatest length — from the posterior surface of the occipital condyles to the
anterior surface of the knob formed by the premaxillae, above the incisors.

2. Condylo-basal length — from the posterior surface of the occipital condyles to
the anterior surface of the premaxillae at the level of the incisors.

3. Snout length — from the most anterior part of the edge of the orbit to the anterior
surface of the premaxillae at the level of the incisors.

4. Antero-posterior width of the zygomatic root of the maxilla — between the
inferior lip of the antorbital foramen and the orbit.

5. Gnathion — is defined by Thomas (1905) as the most anterior point of the pre-
maxillae, on or near the mid line. In Arctocephalus the most anterior point of the
premaxillae is a projecting knob of very variable size, so the gnathion is taken here
as the most anterior surface of the premaxillae at the level of the incisors, and does
not include the knob.

6. Basion — as defined by Thomas (1905), a point in the middle line of the hinder
edge of the basioccipital margin of the foramen magnum.

7. Henselion — as defined by Thomas (1905), the back of the aveolus of either of
the median incisors.

8. Palation — as defined by Thomas (1905), the most anterior point of the hinder
edge of the bony palate, whether in the middle line or on either side of a median spine.

REFERENCES

Albes, E. 1914. The Island of Juan Fernandez. Bull. Pan American Union, 39 : 201-216.
Allen, J. A. 1880. History of North American Pinnipeds. U.S. Geol. and Geog. Survey of
Territories. Misc. Pub. No. 12.

1899. in Jordan, D. S., The Fur Seals and Fur Seal Islands of the North Pacific Ocean.

Washington, pt. III.

1905. The Mammalia of Southern Patagonia. Rep. Princeton Univ. Exp. to Patagonia,

III, pt. I.

Anson, G. 1744. An authentic and genuine journal of Commodore Anson's Expedition. Pub-
lished from the journals of several of the principal officers who attended him throughout
the whole Expedition.

Anthony, A. W. 1925. Expedition to Guadalupe Island, Mexico in 1922. The Birds and
Mammals. Proc. Calif. Acad. Sci. 4th ser., 14 : 277-320.

Bartholomew, G. A. 1950. A male Guadalupe Fur Seal on San Nicolas Island, California.
Journ. Mamm. 31 ; no. 2 ; 175-180.

Bartholomew, G. A., and Hubbs, C. L. 1952. Winter Population of Pinnipeds about
Guadalupe, San Benito and Cedros Islands, Baja California. Journ. Mamm. 33 : no. 2,
160-171.

Baur, G. 1897. Science, N.Y. 5 : 956-957.

Beebe, W. 1924. Galapagos. World's End. xxi -f- 443, 82 figs., 8 pis. New York :
Putnam.

THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA 337

Heller, E. 1904. Mammals of the Galapagos Archipelago, exclusive of the Cetacea. Papers
from the Hopkins Stanford Galapagos Expedition, 1898-99. Proc. Calif. Acad. Sci.
3rd ser., 3 : no. 7, 233-250.

Howay, F. W. 1930. A list of trading vessels in maritime fur trade, 1 785-1 794. Trans.
Roy. Soc. Canada, Toronto, sect. 2, ser. 3, 24 : 111-134.

Korwin, H. 193 i. The Galapagos Islands, or Colon Archipelago. Bull. Pan American
Union, 65 : 11 40-1 145.

Lyon, G. M. 1937. Pinnipeds and a Sea Otter from the Point Mugu Shell Mound of California.
Pub. Univ. Calif. Los Angeles Biol. Sci. 1 : no. 8, 133-168.

Mayer, W. V. 1949. Catalogue of type specimens of Mammals in the Natural History-
Museum of Stanford University. Proc. Calif. Zool. Club, 1 : no. 6, 29-32.

Merriam, C. H. 1897. A new f ur sea -l or sea-bear (Arctocephalus townsendi) from Guadalupe
Island, off Lower California. Proc. Biol. Soc. Washington, 11 : 175-178.

Nehring, A. 1887. Ueber eine Pelzrobben-Art von der Kuste Sud-Brasiliens. Arch, fur
Naturges. 1 : 75-94.

Nybelin, O. 1931. Saugetier- und Vogelcestoden von Juan Fernandez. The Nat. Hist, of
Juan Fernandez and Easter Island. Ed. by Dr. C. Skottsberg. Ill : Zoology, pt. IV,

493-523-
Osgood, W. H. 1943. The Mammals of Chile. Pub. Field Mus. Nat. Hist. Chicago, Zool.

Ser. 30.
Peters, W. 1866. Uber die Ohrenrobben (Seelowen und Seebaren), Otariae, insbesondere

uber die in den Sammlungen zu Berlin befindlichen Arten. Monatsb. Akad. Berlin,

261-281.
— ■ — ■ 1 87 1. Uber eine fur Chile neue Art von Otaria und schloss daran einige Bemerkungen

iiber die Verschiedenheit der in dem atlantischen und stillen Ocean vorkommenden Pelz-

robben. Monatsb. Akad. Berlin, 558-566.
Sivertsen, E. 1953. A new species of Sea-Lion, Zalophus wollebaeki, from the Galapagos

Islands. Det. Kong. Norske Videnskabers Selskabs Forhandlinger, 26 : no. 1, 1-3.
Starks, E. C. 1922. Records of the capture of Fur Seals on land in California. Calif. Fish

and Game, 8 : no. 3, 155-160.
Townsend, C. H. 1924. The Northern Elephant Seal and the Guadalupe Fur Seal. Nat.

Hist. N.Y. 24 : no. 5, 567-578.

1930. Guadalupe Fur Seal in 1929. Bull. N.Y. Zool. Soc. 33 : 32.

1931. The Fur-seal of the Californian Islands. Zoologica, N.Y. 9 : 443-450.

1934- The Fur-Seal of the Galapagos Islands. Ibid. 18 : no. 2, 43-56.

Thomas, O. 1905. Suggestions for the nomenclature of the cranial length measurements
and of the cheek teeth of Mammals. Proc. Biol. Soc. Washington, 18 : 191-196.

Turner, W. 1888. Report on the Seals. Zoology of the Voyage of H.M.S. " Challenger,"
pt. LXVIII.

Wollebaek, A. 1927. El Archipielago de Colon. (Galapagos- 0ene) 11. Dyreliv. M.V.
Norsk Geog. Tidss. 1 : 476-508.

PR r 1SENTED
3 DEC 1954

PLATE 10

a. Ventral view of type skull of Arctocephalus galapagoensis , No. 2812, Stanford University,
California. (Photograph by Dr. Antenor L. de Carvalho.) b. Ventral view of skull of A.
australis, 1880.7.28.7, B.M.(N.H.). c. Lateral view of type skull of A. galapagoensis, No.
2812.

Bull. B.M. (NM.) Zoology. 2, 10

PLATE 10

rwTF"n

PLATE ii

Dorsal views of: a. Type skull of A. galapagoensis , No. 2812. b. Skull of A. australis,
1880.7.28.7, B.M.(N.H.). c. Skull of A. australis, 1949.3. 17. 19, B.M.(N.H.). b and c are
skulls of A. australis of the same condylo-basal length from the Straits of Magellan and the
Falkland Islands respectively.

RESENTED

8 DEC 1954

PRINTED IN GREAT BRITAIN BY
ADLARD AND SON, LIMITED,
BARTHOLOMEW PRESS, DORKING

i 1 JAN 1955

THE SOUTHERN RIGHT
WHALE DOLPHIN,

LISSODELPHIS PERONI (LACEPEDE)

F. C. FRASER

BULLETIN OF
THE BRITISH MUSEUM (NATURAL HISTORY)
ZOOLOGY Vol. 2 No. n

LONDON: 1955

THE SOUTHERN RIGHT WHALE DOLPHIN,
LISSODELPHIS PERONI (LACE'PEDE)

F. C. FRASER

Pp. 339-346 ; PL 12 ; 1 Text-figure

BULLETIN OF

THE BRITISH MUSEUM (NATURAL HISTORY)

ZOOLOGY Vol. 2 No. 11

LONDON: 1955

THE BULLETIN OF THE BRITISH MUSEUM
(NATURAL HISTORY), instituted in 1949, is
issued in five series corresponding to the Departments
of the Museum, and an Historical Series.

Parts appear at irregular intervals as they become
ready. Volumes will contain about three or four
hundred pages, and will not necessarily be compiled
within one calendar year.

This paper is Vol. 2, No. 11 of the Zoological series.

PRINTED BY ORDER OF THE TRUSTEES OF
THE BRITISH MUSEUM

Issued January, 1955. Price Five Shillings.

THE SOUTHERN RIGHT WHALE DOLPHIN,
LISSODELPHIS PERONI (LACEPEDE)

EXTERNAL CHARACTERS AND DISTRIBUTION

By F. C. FRASER

Early in 1952 three dolphins became stranded on the beach at Onekaka, Golden
Bay, which is at the northern end of South Island, New Zealand. Mrs. G. J. Goulter
of Golden Bay took an 8 mm. colour film of the animals, all of which were alive.
They were eventually returned to the sea and swam away apparently none the worse
for their temporary sojourn on dry land. I wish to express my thanks to Mrs.
Goulter for her permission to let prints be made of three of the frames of the film
and also to her brother, Captain R. E. Washbourn, R.N., who brought the film
to the British Museum and helped in the subsequent arrangements for pictures from
it to be used in the present note.

It was clear as soon as the film was projected that the dolphins belonged to the
species Lissodelphis peroni, the Southern Right Whale Dolphin, and although this
species has been recognized for a century and a half I am not aware that any photo-
graphic record of it has previously been published. The photographs (PI. 12)
show the characteristic external features of this species, the most noteworthy of
course being the absence of a dorsal fin and the spectacular coloration of the body.
The pictures of the three animals resting on the sand (PL 12) suggest that they
unlike most dolphins, are broader than deep in the thoracic region. While this
appearance may only be due to the shape assumed by the body when not water-
borne, it is possible that it may be natural and is perhaps a stabilizing factor in an
ocean-going dolphin which lacks a dorsal fin. The slender tail stock can be seen and,
although none of the pictures reproduced here gives a good impression of the pectoral
fins, inspection of the whole film indicated that their shape is, as in very many of
the Delphinidae, with shallowly convex anterior border and with the hinder margin
convex proximally and concave distally. The head is seen to taper to a snout
which, laterally, is well defined from the " forehead " by an angular depression
shallowing as it approaches the middle plane of the head, so that in the profile of the
head the " forehead " merges into the rostrum with only the slightest indication
of a break in the line. The tip of the lower jaw protrudes to a noticeable distance
beyond that of the upper.

The dorsal surface of the body is of a bluish-black colour which is sharply demar-
cated from the glistening white ventral region. Snout, forehead and lower jaws are

ZOOL. 2, II. 19

342 THE SOUTHERN RIGHT WHALE DOLPHIN

white ; the blowhole and eyes come within the pigmented area. From the top of the
head, where the pigment extends anteriorly in a little peak, the boundary between
black and white sweeps low down on the lateral surface of the thorax, reaching its
lowest limit about the insertion of the flippers, which have their origin in the unpig-
mented region. The boundary then takes an upward direction until about half the
depth of the body is reached in the lumbar region. Thence it continues caudad
about midway between the dorsal and ventral margins of the body. A close inspec-
tion of the film was made to decide about the dorsal surface of the tail flukes. Some
indication of the amount of pigment on the tail of one of the specimens can be obtained
from PL 12, fig. 3, but allowance has to be made for the high-light which gives an
impression of a lightness of pigmentation not present in the animal itself. While
in the specimen just referred to the colour was grey-blue, in one at least of the
remaining two the dorsal surface of the flukes was of exactly the same bluish-black
colour as the back. In addition it was seen in the film that the posterior margin
of the white ventral surface of the flukes was bordered by a thin band of dark
pigmentation in one specimen, while in another the band was appreciably wider,
about an inch at its widest. There was definite evidence in part of the film that,
although at first sight the flippers appeared altogether white, at least one of the
animals had a narrow elongated darkly-pigmented patch filling the apex of the
proximal convexity on the posterior border of the limb.

The first published description of L. peroni, that of Lacepede (1804) includes
comment on the contrast of the dark colour of the back with the brilliant white of
the sides and belly and with that at the end of the tail, the extremity of the snout
and on the flippers.

Lesson (1827) gives a more detailed account of the external appearance of this
species of dolphin. " Arrondi dans ses contours, gracieux dans ses formes, lisse
dans toutes ses parties, ce cetace est d'autant plus remarquable, qu'il semble recou-
vert d'un camail noir. Son museau jusqu'a l'oeil est d'un blanc soyeux ou argentin.
II en est de meme des cotes du corps, des nageoires pectorales, du ventre et d'une
partie de la queue. Un large scapulaire, d'un bleu-noir fonce, prenant naissance aux
yeux oil le blanc decrit un croissant, se dessine et se recourbe sur les flancs, pour
recouvrir seulement la partie superieure du dos. Le bord anterieur des nageoires
pectorales et caudales est brun. Le museau est allonge, separe du crane par un
sillon profond." The figure of the specimen (Lesson's PL 9, fig. 1) does not convey
the elegance of form expressed in the description and in particular the shape of the
attenuated snout is quite unlike that of the animals photographed in New Zealand.

Bennett (1840) says : " The upper and hind part of the head, the back, and flukes
are of a uniform deep-black colour which about the lower third of the body, termi-
nates by a straight and abrupt line, leaving the entire abdomen, and inferior portions
of the sides, of a pure and dead white. The snout, and anterior third of the head,
are entirely white ; as also are the swimming paws, with the exception of a broad
black spot on the upper surface and posterior margin of each." Bennett's figure is of
an animal with a disproportionately large head and a body deep in comparison with
its length.

Gray (1846) quotes Cuvier's criticism of the snout shape as depicted by Lesson and

THE SOUTHERN RIGHT WHALE DOLPHIN 343

on his own account draws attention to the restriction in the latter' s figure of the
pigmented dorsal part of the body. He reproduces drawings made at sea which
in these essentials resemble the animals photographed in New Zealand. In Gray's
figures the upper surface of the tail-flukes is of the same dark colour as the dorsal
surface of the body. His lateral view of the dolphin shows the profile of the anterior
end of the lower jaw continuing the line of the profile of the snout. This is not
substantiated in the rounded lower jaw tip of the New Zealand animals.

D'Orbigny and Gervais (1847) have the briefest description of a specimen har-
pooned near Cape Horn. The drawing in their account shews an animal somewhat
slender in form, with the conventional pattern of pigmentation on the dorsal surface
of the body and having the dorsal surface of the flukes of the same black colour.
In addition the dorsal surface of the snout has pigment on it which does not, however,
extend to the margin of the upper jaw and stops short of the pigmentation on the
top of the head. The flipper is shown bearing a large irregular black spot about
the middle of the posterior edge, and on the tail stock, just anterior to the insertion
of the flukes, four short oblique flecks of dark colour project ventro-caudally from
the boundary which separates the black of the back from the white of the under
surface.

Philippi (1893) describes two specimens caught off the east of Patagonia about
41 S. The figure which he gives of one of these indicates that, in general, the animal
was like those described by earlier writers. Philippi says that the flippers were
pure white above and below, only the rearmost corner and tip being black, and the
tail black above, white below with black posterior border.

Lillie (1915) refers briefly to L. peroni having been seen twice, a pair of animals
on each occasion. He states that all four were exactly alike and agreed with Gray's
figure except that the tail flukes were quite white above and below.

Summing up the foregoing descriptions, it would appear likely that only one species
of Lissodelphis is involved for which the name L. peroni has priority. The colour
pattern is constant so far as it concerns the amount of black on the back and white
on the belly but varies in the amount of pigmentation on the snout, flippers and on
the dorsal surface and ventral margin of the tail flukes.

This colour pattern is clearly distinguishable from that of the northern form
L. borealis (Peale). Excellent photographs of this northern species are to be found
in the account of Scheffer and Slipp (1948). These show an animal which is much
more extensively pigmented than its southern congener. In L. borealis the dark
pigmentation covers the greater part of the body. The rostrum and anterior part
of the " forehead " appear light grey in one of the photographs in the paper just
referred to and the end of the lower jaw is without pigment. A roughly lozenge-
shaped area of white lies between the flippers on the ventral surface of the body
which, anteriorly, extends on to the throat, where it ends in a sharp angle. Pos-
teriorly it is drawn out to become a thin white line in the umbilical region and thence
to the tail it widens out again very gradually. The dorsal surface of the tail is black
and the ventral surface white with a black band on the middle line and posterior
border.

In the consideration of the geographical distribution of L. peroni it is convenient

344 THE SOUTHERN RIGHT WHALE DOLPHIN

to start with the type locality and continue eastwards round the globe. Peron
(1807) himself says that the species which bears his name was based on a specimen
caught off the south coast of Tasmania ; on the day following the capture the ship's
position was 44 S., 141 ° 27' E. Flower and Garson (1884) record a skull which was
sent to the Royal College of Surgeons from Tasmania. Lillie saw the dolphin on
two occasions in positions 42 51' S., 153 56' E. and 47 04' S., 171 33' E. respectively,
the former being eastward of Tasmania, the latter off the southern end of the east
coast of South Island, New Zealand.

The three specimens which are the subject of the present paper stranded on the
north end of South Island. The specimen figured by Gray (1846) was seen in position
46 48' S., 142 W. that is roughly halfway between New Zealand and the southern
end of South America.

In the neighbourhood of Cape Horn, Malm's (1871) record of a skull from " Southern
Chile " is not by itself very indicative as to exact locality but it may legitimately be
mentioned in association with the other reports from the same region. Lesson (1827)
saw the dolphin near Pillar Cape at the western end of the Straits of Magellan and
says that on the 12th January, 1823, several hundred surrounded the " Coquille ",
the ship then being three days south of the Chonos Archipelago, Chile, that is about
45 S. Bennett (1840) first saw specimens of L. peroni in 40 S., 50 W., south
eastwards from the mouth of the Rio de la Plata. It was afterwards frequently
seen by him during the passage round the Horn as far south as 54 S., but was not
observed in lower latitudes than 40 S. on the western side of Cape Horn, nor indeed
during any subsequent part of the voyage. Philippi's (1893) specimens, a male and
a female, were caught off the east coast of Patagonia about 41 S. D'Orbigny and
Gervais (1847) say it was observed around Cape Horn from 58 to 64 S. Mr. R. M.
Laws (St. Catherine's College, Cambridge) to whom I am indebted for the record,
has informed me that while on a passage from Monte Video to Stanley, Falkland
Islands, he saw dolphins which were unquestionably L. peroni. The extract from
his diary reads : " Jan. 22nd, 1948, 16*50 hours. A school of about 60 dolphins
approached the ship from a S.E. direction. Some accompanied us for a few
minutes. . . . There were two types, one greyish on top and white below with no
definite line of demarcation and with a very pointed beak. It also had two parallel
white lines along the flanks. (Probably Lagenorhynchus obscurus. F.C.F.). The
other species was black and white clear cut, with no dorsal fin and having a white
beak ... a fix at 19.40 hours gave our position as 42 2' S., 56 6' W."

Eastwards across the South Atlantic at the southern end of the African continent
two records are known. Schlegel (1841) refers to a skull which came from the
Cape of Good Hope, but True (1889) is doubtful about the identity of this specimen.
The other record is that of Dr. J. E. Hamilton, who in 1927 recorded and made a
drawing of a dolphin which is certainly L. peroni, seen in position 38 34' S., 8° 06' E.

The only remaining acceptable record, and it is not a very precise one, is that of
Lesson (loc. cit.), who says that the dolphin " was seen in 45 S. in circumnavigating
Australia."

Quoy and Gaimard (1824) refer to a dolphin with long white beak seen in 2 S.
near New Guinea. This was distinguished as " delphinus Peroni de Lacepede ".

THE SOUTHERN RIGHT WHALE DOLPHIN

345

Neither this brief description nor the place of occurrence justifies acceptance of the
record.

The well-substantiated places of occurrence of Lissodelphis peroni indicate that
it probably ranges round the world in the southern hemisphere. Although it is not
entirely restricted to the West Wind Drift it appears to have some predilection for it,
because, as shown in Fig. 4, those records to the north of the Sub-tropical Converg-
ence are for the most part close to that boundary, and except for one sight record
there is no evidence that it penetrates into the Antarctic Ocean.

Text-fig. i. Distribution of Lissodelphis peroni. Complete circles indicate precise
positions, dotted circles and lines indicate less well denned places of occurrence.

346

THE SOUTHERN RIGHT WHALE DOLPHIN

BIBLIOGRAPHY

Bennett, F. D. 1840. A Whaling Voyage round the Globe. 2 : 1-395, figures in text. London.
D'Orbigny, A., & Gervais, P. 1847. Voyage dans I'Amerique Meridionale. 4 : Mammiferes.

1-32, Atlas, PI. XXL
Flower, W. H., & Garson, J. G. 1884. Catalogue of the Specimens ... in the Royal College

of Surgeons of England. Pt. 2. Mammalia other than Man. London (J. & A. Churchill).
Gray, J. E. 1846. The Zoology of the Voyage of H.M.S. " Erebus and Terror," i83g~43.

Vol. 1. Mammalia, Birds. 13-53, 37 pi-
Lacepjede, B. 1804. Histoire Naturelle des Cetaces. xliv + 329, 16 pi. Paris,
Lesson, R. P. 1827. Voyage autour du Monde ... sur . . . "La Coquille." Zool. 1:

iv -f- 743, 50 pi. in atlas. Paris (Arthur Bertrand).
Lillie, D. G. 1915. Cetacea. " Terra Nova " Reports Zool. 1 : No. 3, 85-124, 14 figs, 8 pi.

London. Trustees of Brit. Mus.
Malm, A. W. 1871. Hvaldjur i Sveriges Museer, Ar 1869. K. Svenska. Vetensk. Akad.

Handl. 9 : no. 2, 1-104, 6. pi.
Peron, F. 1807. Voyage de Decouvertes aux Terres Australes, 1 : xv + 498. Paris.
Philippi, R. A. 1893. Die Delphine an der Siidspitze Sudamerikas. Ann. Mus. Nac.

Chile : 1-16, 5 pi.
Quoy, J. R. C, & Gaimard, P. 1824. Voyage autour du Monde ... " Oranie " et " Physi-

cienne ". Zool. : 1-671. Paris (Pillet).
Scheffer, V. B., & Slipp, J. W. 1 948. The Whales and Dolphins of Washington State with

a Key to the Cetaceans of the West Coast of North America. Amer. Mid. Nat. 39 : No. 2,

257-337, 5o figs.
Schlegel, H. 1841. Abhandlungen aus dem Gebiete der Zoologie und vergleichenden Anatomie.

Heft. 1, 1-44, 6 pi. Leiden.
True, F. W. 1889. Contributions to the Natural History of the Cetaceans. A Review of

the Family Delphinidae. Bull. U. S. Nat. Mus. 36 : 1-191, 47 pi.

EXPLANATION OF PLATE
PLATE 12

Figs. 1-3. Lissodelphis peroni. Figs. 1 and 2. The dolphins on the beach at
Onekaka, South Island, New Zealand. Fig. 3. One of the animals having been
returned to the sea but not water-borne. [Photo : Mrs. G. L. Goulter).

/^STHZ?^

PR!

i 1 JAN 1955

Bull. B.M. (N.H.) Zool. II, n.

PLATE i

LISSODELPHIS PERONI

54L

PRESENTED

3 1 JAN 1955

PRINTED IN GREAT BRITAIN BY
ADLARD AND SON, LIMITED,
BARTHOLOMEW PRESS, DORKING

NEW SPECIES OF TILAPIA

(PISCES, CICHLIDAE) FROM

LAKE JIPE AND THE

PANGANI RIVER,

EAST AFRICA

ROSEMARY H. LOWE

BULLETIN OP
THE BRITISH MUSEUM (NATURAL HISTORY)
ZOOLOGY Vol. 2 No. 12

LONDON: 1955

NEW SPECIES OF TIL API A (PISCES, CICHLIDAE)
FROM LAKE J1PE AND THE PANGANI RIVER,

EAST AFRICA

ROSEMARY H. LOWE

(East African Fisheries Research Laboratory, Jinja, Uganda)

Pp. 347-368 ; Pis. 13-17 ; 4 Text-figures

BULLETIN OF

THE BRITISH MUSEUM (NATURAL HISTORY)

ZOOLOGY Vol. 2 No. 12

LONDON: 1955

THE BULLETIN OF THE BRITISH MUSEUM
(NATURAL HISTORY), instituted, in 1949, is
issued in five series corresponding to the Departments
of the Museum, and an Historical series.

Parts will appear at irregular intervals as they become
ready. Volumes will contain about three or four
hundred pages, and will not necessarily be completed
within one calendar year.

This paper is Vol. 2, No. 12 of the Zoological series.

PRINTED BY ORDER OF THE TRUSTEES OF
THE BRITISH MUSEUM

Issued June, 1955 Price Ten Shillings

NEW SPECIES OF TILAPIA (PISCES, CICHLIDAE)
FROM LAKE JIPE AND THE PANGANI RIVER,

EAST AFRICA

ROSEMARY H. LOWE

CONTENTS

I. Introduction .......... 347

II. Systematic Diagnosis ........ 352

III. Field and Pond Observations ....... 356

a. T. girigan sp. n and T. jipe sp. n. ...... 356

b. T. pangani sp. n . . . . . . . . 361

c. T. mossambica korogwe subsp. n. . . . . . . . 361

d. Lake Chala Tilapia, T. hunteri Giinther ..... 362

IV. Lake Victoria Species of Tilapia Grown in Ponds . . . 362
V. Discussion ........... 363

VI. Summary ........... 365

Acknowledgements ......... 366

References .......... 366

SYNOPSIS

Three new species and one new subspecies of Tilapia are described from the Pangani River
system. Field and pond observations are given, and also observations on the growth of
T. esculenta Graham and T. variabilis Boulenger, collected as fry from Lake Victoria and reared
in ponds.

INTRODUCTION

In recent years recognition of the need for more protein in the diet of Africans
has led to investigations of the growth of African fish in ponds. Certain species
of Cichlid fishes of the genus Tilapia have proved very amenable to pond culture.
This paper is concerned with an investigation, made in January, 195 1, of the Tilapia
species of the Pangani River system, certain of which were already being grown
in fish ponds.

The Pangani River flows through Tanganyika Territory in a south-easterly
direction from Lake Jipe, which lies on the Kenya border just south-east of Kili-
manjaro, to the Indian Ocean (Text-fig. 1). Samples of Tilapia were collected
from Lake Jipe, from ponds at Taveta near Lake Jipe, and from the Tanganyika
Government's experimental fish farm at Korogwe. The Taveta ponds were stocked
with fry from Lake Jipe. The Korogwe ponds were supplied with water by a
channel leading from the Pangani River, and separate ponds were stocked with

ZOOL. II, I?, 20

35o

NEW SPECIES OF TILAPIA FROM EAST AFRICA

Pangani River Tilapia and Lake Jipe Tilapia; Lake Victoria and Belgian Congo
species foreign to the Pangani system had been introduced into other ponds at
Korogwe.

Lake Jipe is connected with the Ruvu or Pangani River in wet seasons through
a series of swamps. In 195 1 the lake was about twelve miles long by one and-a-half
miles wide, but was said to be silting up and spreading ; it was a shallow lake, only
a few feet deep over much of its area, with large patches of water plants, Potamogeton
and Najas species, rising to the surface, and frequented by numerous water birds.

KENYA

Text-fig. i. The Pangani River system, Tanganyika Territory.

NEW SPECIES OF TILAPIA FROM EAST AFRICA 351

The Tilapia collected were examined at the British Museum (Natural History)
together with other specimens from Lake Jipe collected previously by Mr. H. Copley,
Fish Warden, Kenya, and from the Pangani River collected by Major R. E. Gould,
Fisheries Officer i/c Fish Farming in Tanganyika Territory. The following Tilapias
considered to be new and described below were found:

T. jipe sp. n. and T. girigan sp. n. in Lake Jipe ;

T. pangani sp. n. from the Pangani river at Korogwe ; and

T. mossambica korogwe subsp. n. from Korogwe ponds fed by the Pangani

River.

Although, at first, resemblances to T. mossambica Peters and T. nigra Giinther
suggested that the Jipe and Pangani forms might find a place with these in one
supraspecies, closer examination has emphasized the differences and, except for T.
mossambica korogwe subsp. n., they are here described as distinct species, with some
reservation as to the rank of the allopatric T. pangani and T. girigan relative to
each other. T. jipe though superficially like T. girigan has very distinctive pharyn-
geal teeth quite unlike T. mossambica and more like T. galilaea (Linn.). Further
knowledge of the ecology of all these species is required before the interspecific
relationships can be understood.

These Tilapia are, however, already being used on fish farms and transported
round the countryside. It is therefore very desirable that those responsible for the
fish should be able to name them to record movements, quite apart from the desira-
bility of recording as much as possible about natural distribution before species from
different river systems are inevitably mixed. For example, further study has shown
that fish of supposed "T. nilotica" grown in the Tanganyika Government's experi-
mental fish farm at Korogwe really belong to the new species T. pangani, and
appear to be more nearly related to T. mossambica than T. nilotica ; this explains
the differences in behaviour between these fish and the genuine T. nilotica which
have been tried in dams in Uganda and the Belgian Congo.

The situation has been further complicated by the introduction of four distinct
species foreign to the Pangani system into the Korogwe ponds, T. esculenta Graham
and T. variabilis Blgr. from Lake Victoria, and T. macrochir Blgr. and T. melanopleura
A. Dum., from the Belgian Congo. There are also reports that T. nigra Giinther was
once introduced into streams flowing south from Kilimanjaro, streams which also
find their way into the Pangani.

T. mossambica korogwe which was not intentionally stocked in the Korogwe ponds
was nevertheless found in them and presumably gained access from the Pangani
water supply together with some small Haplochromis which also appeared in these
ponds. These accidental entries into ponds indicate the difficulty of making really
fishproof screens and the need to examine the indigenous Tilapias in these rivers
before escapes from ponds, and possibly hybridization with the local species, occur.

A summary of the diagnostic characters of the new forms is given in Table I,
and some of their characters are compared with those of T, mossambica, T. nigra
and 7\ nilotica in Text-fig. 2.

352

NEW SPECIES OF TILAPIA FROM EAST AFRICA

t-»

1 *>

1 CO

I
o

•

vO On fOCO

•ri

tUD

•^

u->

-<f

CO
CO

CO
tN

«>.

->*■

•**

■*

St)

fei

4«

<tt

•tSk

>^-

l>.

On ir>

CO M

«*-

«t-

CO c

r r

o ^-

tfn

"tt>£

? "So

« ^

^ fl

4->
0,

>, W)

*.a

^ £

■2*

o ,2}

T5 rt

•a g

£ o

C CD

* P^

NEW SPECIES OF TILAPIA FROM EAST AFRICA

353

■a-S *

-CJ

r/)

</)

•o

I I I I I

1 1

I ii

: = •

1 s

o £

£< d

cd co
OT ft

<D .

"S #

CD s_^

CO -«s>

o d
d rt

<3«!

o S3 n

5 O N

^•c «

=3 d o

'So* "
«+*

O °00
d i- M

d £

r^ K

S 8 s

h. fc h

O en

«1

a. d
fi °
§ w

t?I

1 co id

«£ d
• rf ^

E« 2
ESS

H
w
H

354 NEW SPECIES OF TILAPIA FROM EAST AFRICA

II. SYSTEMATIC DIAGNOSES

Tilapia jipe sp. n.
(PI. 13 and i6a)

16 specimens of 5-20 cm. total length from Lake Jipe.

54 specimens of 7-28 cm. from ponds stocked from Lake Jipe.

Tilapia with body proportions, teeth, fin rays, scales and gill-rakers as in the
accompanying table of diagnostic characters (Table I) . Lower pharyngeal bone with
heart-shaped dentigerous area covered with dense velvet of fine teeth (PL i6a).
Dorsal profile of head convex in smaller specimens but jaws become elongated in
mature males giving a concave upper profile. Fin rays long, pectoral generally
reaching origin of anal and posterior tip of dorsal extending halfway along caudal ;
caudal fin long with very definite vertical black or dark brown stripes.

Non-breeding fish light grey with spotted effect caused by the very distinct black
or dark brown centre to each scale along back and sides, belly light grey. Faint
indications of 7-9 darker vertical stripes on body of some specimens; dorsal and anal
fins grey, spotted anteriorly and striped posteriorly with black ; caudal fins with
vertical stripes ; pectoral fins unpigmented, pelvics becoming dark in larger fish.
Breeding fish, particularly males, develop green and purple sheen, bright orange
edge to dorsal fin and crimson edge to caudal, bluish-green pelvics and bluish-green
edge to anal in some specimens ; no genital tassel. Fry up to about 8 cm. long have
olive-greenish body with 10-14 thin vertical dark brown stripes stretching from
dorsal to ventral surface, caudal and hind end of anal striped as in adults ; well
marked black " Tilapia mark " on a clear background extending from the last
dorsal spine to the third or fourth soft ray, and followed by three definite black bars
on hind end of dorsal fin.

Among the 70 T. jipe preserved, 12 (17%) had three anal spines, 5 (7%) had
three plus a partly ossified fourth, 52 (75%) had four and 1 (1%) had five.

The type is a male of 162 + 45 mm. (British Museum (Natural History) register
number 1952.2.26.2) and the allotype is a female of 140 + 35 mm. (B.M.(N.H.),
1952.2.26. 1), both collected from Lake Jipe in January, 1951.

The striped tail (previously thought to be characteristic of T. nilotica in East
Africa) and general appearance of T. jipe are similar to T. nilotica Linn., but in
contrast with this species, T. jipe generally has 4 anal spines and a shallower body,
smaller head, wider interorbital, more rows of teeth in the jaws, and the jaws of
mature males become accentuated, a characteristic of the mossambica group of
Tilapia and not shown by T. nilotica. The pharyngeal bone of T. jipe is quite distinct
and characteristic, the dense velvet of fine pharyngeal teeth being most like that of
T. galilaea (L.) ; however, T. jipe differs from T. galilaea in having a shallower
body, a longer snout accentuated in mature males, a less deep preorbital bone, more
lateral line scales, and usually four anal spines.

T. jipe is easily distinguished from the sympatric T. girigan by the striped
caudal fin, four anal spines, pectoral fin reaching origin of anal fin, the heart-shaped

NEW SPECIES OF TILAPIA FROM EAST AFRICA 355

dentigerous area and fine teeth on the lower pharyngeal bone, less concave profile,
and lighter body colour with very distinct spots.

T. girigan sp.n.
(PL 14 and i6b)

22 specimens of 5-31*5 cm. total length from Lake Jipe.

9 specimens of 7-23 cm. total length from ponds stocked from Lake Jipe.

Tilapia with body proportions, teeth, fin rays, scales and gill-rakers as shown in
the diagnostic Table I. Lower pharyngeal bone with large triangular toothed area
and straight posterior edge ; pharyngeal teeth coarse and well scattered on bone
(PL i6b). Jaws becoming elongated in mature fish, more markedly in the males,
giving concave upper profile. Fin rays not very long, pectoral not reaching origin
of anal; caudal fin immaculate or uniformly dark, or with spots and blotches, but not
striped.

Non-breeding fish dark olive-brown, spotted effect due to black or dark brown
centre to all the scales on back and sides extending right down to the grey ventral
surface. Indications of about nine vertical black or dark brown stripes on some
specimens. Dorsal and anal fins dark grey with posterior ends spotted or blotched
with black ; pectorals immaculate ; pelvics dark grey. Breeding males with red or
orange margin to dorsal and caudal fins; no genital tassel. Fry up to about 8 cm.
total length olive-brown with 10-14 vertical dark stripes; " Tilapia mark " as in
T. jipe fry, but general body colour darker and caudal and hind end of dorsal imma-
culate or spotted as in adults, but not striped as in T. jipe fry.

The type is a male of 210 + 49 mm. (British Museum (Natural History) register
number 1952.2.26.3), and the allotype is a female of 177 + 45 mm. (B.M.(N.H.),
1952.2.26.4), both collected from Lake Jipe in January, 1951.

In general body shape and in the development of the jaws and the concave profile
of the mature males T. girigan resembles T. mossambica Peters. It differs from
T. mossambica, however, in having a higher number of dorsal and anal fin rays [30-31
dorsal rays instead of (26) 27-29 (30), and 14-16 anal rays instead of 12-15], more
scales in the lateral line series (33-36) instead of 30-32 in T. mossambica), more
gillrakers [(17) 18-20 in contrast to 16-19 ( 20 )L an d the outer teeth in the jaws
remain bicuspid, whereas they often become unicuspid in mature male T. mossambica.
Also the lower pharyngeal bone of T. girigan has a shorter blade and stouter appear-
ance than in T. mossambica (compare PL 16. b and E.). The pharyngeal teeth of
T. girigan are most like T. nilotica Linn., (PL i6f) but the toothed area is larger in
T. girigan. T. girigan has a higher number of scales in the lateral line series (33-
36 instead of 31-33 in T. nilotica), and lacks the striped tail of T. nilotica. Males
of T. girigan show lengthening of the jaws and development of a concave profile
as in T. mossambica, a character not found in T. nilotica. T. girigan is very like
T. pangani and may have arisen as a lake form of this species, but in T. girigan the
lower pharyngeal bone and the toothed area are larger and the teeth considerably
coarser, longer, stronger and more scattered on the bone than in T. pangani of com-
parable sizes, and the preorbital bone is not so deep. T. girigan from Lake Jipe

ZOOL. II, 12. 21

356 NEW SPECIES OF TILAPIA FROM EAST AFRICA

were considerably darker in colour than T. jipe from the lake, and darker and more
spotted than T. pangani ; in both T. girigan and T. jipe specimens from the ponds
were lighter in colour and more spotted than those from the lake.

T. pangani sp. n.
(PL 15 and 16c).

8 specimens of 22-29 cm. total length.

Tilapia with body proportions, teeth, fin rays, scales and gillrakers as shown in the
diagnostic Table I. Lower pharyngeal bone smaller and with smaller toothed area,
and with a flatter posterior edge than that of T. girigan ; pharyngeal teeth finer,
particularly anteriorly, and closer together than in T. girigan, but coarser than in
T. jipe (compare PL 16c with A and b) . Jaws lengthening in mature males, giving
concave upper profile. Dorsal, anal and caudal fin rays as long as in T. girigan ;
pectoral not reaching origin of anal; caudal fin with dark spots or blotches, or slight
stripes at base only (not striped as in T. jipe) .

Non-breeding fish brownish or grey, dark centres to scales on body but general
impression not so spotted as in T. girigan or T. jipe ; indications of vertical stripes
(about 9) in some specimens; pectoral fins, and in some small ones pelvics, imma-
culate; pelvics pigmented in larger specimens. Posterior end of dorsal and anal fins
spotted or barred with dark brown or dark grey. Breeding fish, particularly males,
lose the vertical stripes, develop a purple-green sheen, dark pelvic fins, bright
orange margin to dorsal and bright crimson margin to caudal fin ; no genital tassel.

The type is a male of 236 -f 74 mm. collected from the Pangani River near Korogwe
by Major Gould in 1949 [British Museum (Natural History), register number 1952.2.
26.39], and the allotype is a female of 189+46 mm. collected from a Korogwe
fishpond, stocked from the Pangani River, in January, 1951 [B.M.(N.H.), 1952.
2.26.40].

T. pangani is like T. mossambica Peters in general facies and development of the
jaws and concave profile of the mature male, but in contrast with T. mossambica
it has (a) a higher number of dorsal fin rays [30-31 ; contrast (26) 27-29 (30)] ; (b)
a higher number of anal rays (14-16 contrast ; 12-15) ; (c) more scales in the lateral
line series [(33) 34-35 ; contrast 30-32]. T. pangani is distinguishable from T.
girigan by the more mossambica-like pharyngeal teeth on the smaller lower pharyngeal
bone, and the deeper preorbital bone.

It is interesting to note here a report from Major Gould that whereas T. mos-
sambica from the lower Pangani can be transferred to, and live in, seawater lagoons,
T. pangani does not survive this transfer.

T. mossambica korogwe subsp. n.

(PL i6d).

12 specimens of 11-18 cm. total length.

Tilapia with body proportions, teeth, fin rays, scales and gillrakers as shown in
the diagnostic Table I. Lower pharyngeal bone with short blade, toothed area very

NEW SPECIES OF TILAPIA FROM EAST AFRICA

357

like T. mossambica but teeth slightly coarser. Mouth very large, wide and thick-
lipped with wide tooth bands; jaw teeth small, poorly developed, weakly cuspidate
and well embedded, teeth of both outer and inner series tending to be unicuspid at
the sides of the mouth in almost all specimens (including the female). Eyes much
more dorso-lateral than in other Tilapia. Pectoral fins short, not reaching anus.

Body dark silvery grey, longitudinal black bands along lower lateral line, with
indications of about 7 vertical black bars across this line. Line of 4 or 5 black spots
above dorsal lateral line and black spot on dorsal side of caudal peduncle. "Tilapia
mark " an oval black spot near hind end of dorsal fin. Tail immaculate dark grey.
Long dorsal and anal fins ; anal and pelvics and ventral surface of body dark grey.
Mature males with white or grey tip to caudal fin and small (2 mm.) white genital
papilla. Female (one specimen only) lighter grey with more definite vertical
black stripes (8) and black spot on dorsal side of caudal peduncle ; mouth not so
large as in males.

This Tilapia appeared in several of the fishponds at the Tanganyika Government
experimental fish farm at Korogwe. Its origin is uncertain but it seems most pro-
bable that it gained access to the ponds with the water supply, which is a cut flowing
from, and returning to, the Pangani River. Only one female, caught together with
the males and believed to be the female of this subspecies, was found, so the descrip-
tion of the female given above must be considered with reservation.

The type is a male of 140 -f- 43 mm. collected by Gould in 1951, from one of the
Korogwe ponds [British Museum (Natural History) register number 1952.2.26.37].
The one female of 91 -f 29 mm. collected in January, 195 1, from one of the Korogwe
ponds [B.M.(N.EL), 1952.2.26.5] is considered, on present evidence, to be the
female of this subspecies and is, therefore, selected as the allotype.

T. mossambica korogwe is very like T. m. mossambica in general facies and in the
development of the concave upper profile in the males and in the simplification
of the jaw teeth, but it has a lower number of gillrakers [14-17 instead of 16-19 (20)
in T. mossambica], a higher number of dorsal spines (16-18; contrast 15-17 in T.
mossambica), and four anal spines in some specimens.

In T. mossambica korogwe the lower pharyngeal bone has a slightly shorter blade
and the pharyngeal teeth are slightly coarser than in T. mossambica. The head and
mouth are wider in T. mossambica korogwe and the simplification of the jaw teeth is
even more marked than in T. mossambica of comparable sizes and extends to the
inner series of teeth. T. mossambica korogwe is easily distinguished from T. jipe,
T. girigan, T. pangani, T. nilotica and T. esculenta by the following characters :
the very wide head, wide mouth and wide lips, relatively narrow interorbital and the
dorso-lateral eyes, small eyes, large preorbital, elongated very concave snout, pelvics
far forward and pelvic-anus distance long.

OTHER TILAPIA FROM THE PANGANI RIVER

The British Museum collection contains three other specimens of Tilapia from the
Pangani River and certain fry collected from the Korogwe fish ponds which do not
seem to belong to T. pangani, T. mossambica korogwe or any of the Tilapia species

358 NEW SPECIES OF TILAPIA FROM EAST AFRICA

stocked in the ponds. Two of these specimens were collected by Gould in 1949 from
the Pangani River together with T. pangani ; the male was much darker than
T. pangani, less spotted and with indications of 7 vertical dark bars. These two
fishes differed from T. pangani in having (a) 4 anal spines and 10 soft rays, (b) 17-18
dorsal spines and 10 soft rays, (c) 31 scales in the lateral line series (instead of 33-35),
(d) 15-16 gillrakers on the lower part of the anterior arch (contrast 19-21 in
T. pangani). In all these characters they correspond well with T. nigra Gthr.
There are reports that T. nigra were once introduced into the streams flowing from
the south side of Kilimanjaro ; such streams drain into the Pangani, so these fishes
may not be indigenous. Two small Tilapia (7-8 cm.) found in the Korogwe ponds
stocked with Tilapia from the Taveta ponds also appear to be T. nigra, and may have
entered this pond with the water supply from the Pangani.

The British Museum collection also contains a 17 cm. Tilapia with three anal spines,
33 scales in the lateral line series and 20 gill rakers, collected from the Pangani
by Playfair in 1865, and identified as Chromis niloticus 1 (= T. nilotica). This fish
does not appear to belong to any of the other species described from the Pangani
system. Chromis niloticus (Hasselquist) collected from the Ruvu (Pangani) at
Korogwe are also described by Pfeffer (1896), but the number of specimens seen is not
stated. Boulenger (1915) assigns the C. niloticus described by Pfeffer to Tilapia
natalensis (Peters) and T. nilotica L. The illustration given by Pfeffer shows that
in general facies his C. niloticus differs considerably from T. nilotica from the type
locality, although it agrees with T. nilotica in many characters such as the numbers of
lateral line scales and gillrakers. No other specimens like Playfair's fish have been
found in recent collections; until further material is forthcoming it is impossible
to say how this and Pfeffer's fishes relate to the other Tilapias.

III. FIELD AND POND OBSERVATION

A. T. girigan and T. jipe in Lake Jipe and ponds.

Prior to these investigations it was presumed that the Tilapia population in Lake
Jipe consisted of one or possibly two new species related to T. mossambica. No
specific name had yet been given to these fish which were merely referred to as
Lake Jipe Tilapia. Field observations did not reveal for certain the existence of
two species, but it was noticed that some fishes had four and some had three anal
spines, and that the four-spined fishes had a striped caudal fin whereas in the three-
spined the caudal fin was striped, spotted or immaculate.

Subsequent examination of samples kept for laboratory study showed that all
the four-spined Tilapia belonged to the species now called T. jipe, whereas the three-
spined Tilapia included both T. jipe and T. girigan, in the proportion 1 . 12 T. jipe :
1. 00 T. girigan. The proportions of the two species in the preserved samples
suggested that most of the Tilapia seen from Lake Jipe happened to be T. girigan,
whereas most of those from the ponds were T. jipe. Many more T. jipe than T.
girigan were found among the fry used for stocking the ponds, which helps to explain

1 Playfair & Guijther, 1896, page iij,

NEW SPECIES OF TILAPIA FROM EAST AFRICA

359

the predominance of T. jipe in the ponds ; these fry, 2-10 cm. long, were caught by-
pulling a sack-cloth seine to the shore through a clearing in the reeds bordering Lake
Jipe. Whether T. girigan fry inhabit a different zone in the lake or would be found
in greater numbers at other times of the year is not known. The large Tilapia
from Lake Jipe, among which T. girigan predominated, were caught by beating the
water around gillnets set by day among the patches of water weed. 1

Twenty-two Lake Jipe Tilapia were taken from Taveta to Korogwe Ponds,
sixteen had four anal spines and were probably T. jipe, three had three anal spines
and may have included T. girigan, and three had a partly ossified fourth spine.
Two of these Tilapia were presented to the British Museum in December, 1951, and
identified as T. jipe ; this confirms the presence of this species at Korogwe. The
presence of T. girigan at Korogwe needs confirmation ; a sample of fry from Korogwe
ponds indicates that T. girigan were also there, but T. girigan fry are difficult to
distinguish from T. pangani fry which might have found their way into the ponds
accidentally with the water supply.

Length frequencies of male and female T. jipe and T. girigan are given in Text-
fig. 3 a and B and summarized in Table II.

Table II. — The sex ratio, length range and minimum breeding sizes of T. girigan and
T. jipe.

Minimum

Sex ratio

Body length (cm.)

breeding size

Fishing

( '

( "

Place

method

Species

Male

Female

Male

Female

Male

Female

L. Jipe .

Gillnet

Mainly T. .
girigan

26 .

16-32

r T. girigan and .

143

14-28

15-22

Taveta pond .

Pond
emptied

T. jipe (3 anal

spines)
T. jipe (4 anal .

spines)

216

36 .

15-25

14-21

Taveta pond •

Angle

Mainly T. jipe .

17-27

15-17

—

Korogwe

Pond

T. jipe and .

14 .

23-3°

16-23

—

pond

emptied

? T. girigan

It is immediately clear that in both lake and pond samples, and in those in which
T. girigan or T. jipe predominated :

(a) the males were considerably larger than the females ;

(b) females started to spawn at a smaller size than the males ;

(d) the pond reared fish were of comparable lengths and minimum breeding
sizes with the fish from the lake.

The length /weight relations of T. girigan and T. jipe are indicated in Text-fig. 4.
The pond-reared fish were in very poor condition, i.e. showed a low weight for length,
compared with the Tilapia from Lake Jipe. This was evidently the result of lack
of food in the ponds.

1 Peter sius tangensis Lonnberg, previously recorded from Tanga, and Barbus paludinosus Peters, a
species widely distributed in East Africa, were also caught in Lake Jipe. These species grow only to
about 10 cm. long. A few Clarias mossambicus Peters, another species widely distributed in East
Africa, were found when the Taveta pond was emptied.

360

NEW SPECIES OF TILAPIA FROM EAST AFRICA

The Taveta pond which was emptied on 16th January, 195 1, was said to have been
stocked with 4-8 cm. Tilapia from Lake Jipe about fifteen months previously. After
fifteen months in the pond the T. jipe (i.e. those Tilapia with four anal spines)
males were 15-25 cm. long (about 40-200 gm.) with pronounced modes at 17 cm.

35-

30 ^

3 20'

X
u.

<£X

10 12 14 16 18 20 22 24
TOTAL BODY LENGTH -CM

30 32 34

Text-fig. 3. Length frequencies of Lake Jipe and Pangani River Tilapia reared in fish-
ponds. Males , females ■ — . a. Tilapia with four anal spines,

probably all T. jipe, examined when a Taveta pond was emptied fifteen months after
stocking with 2-10 cm. fish from Lake Jipe. b. Tilapia from Lake Jipe mainly T,
girigan. c. T. pangani from a Korogwe pond,

NEW SPECIES OF TILAPIA FROM EAST AFRICA

361

500

480

460

440

420

400

380

360

340

320

300

280

260

240 \

220

200

180

160-

140.

120.

100

80 4

40 4

rf^T^7

2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34

TOTAL BODY LENGTH- CM

Text-fig. 4. The length/weight relationship of Lake Jipe Tilapia. Data from 27
Tilapia from Lake Jipe, mainly T. girigan, marked thus x ; all other Tilapia from
Taveta fishponds and mainly T. jipe. (Curve fitted by eye.)

362 NEW SPECIES OF TILAPIA FROM EAST AFRICA

and 22-23 cm., and females were 14-21 cm. (about 30-120 gm.), with a pronounced
mode at 17 cm. (Text. fig. 3A). Among the three-anal-spined Tilapia (T. girigan
and T. jipe) from the pond the males were 14-28 cm. (30-206 gm.) and females
15-22 cm. (40-140 gm.). Both species had been breeding for some time as the pond
contained numerous fry of 6-1 1 cm.; T. jipe fry were the more numerous but fry of
both species were found.

Adult Tilapia from Taveta ponds were taken to Korogwe in October, 1950, and
numerous fry 2-6 cm. long of T. jipe and probably also T. girigan (see above) were
seen when the pond was examined in January, 195 1.

Text-fig 3B gives the relative sizes of males and females from catches in Lake Jipe
in which T. girigan were highly predominant. Both T. girigan and T. jipe females
were breeding at a considerably smaller size than the males. In Lake Jipe T. girigan
males of 21 cm. (about 140 gm.) and females of 17 cm. (about 60 gm.) were spawning.
In the Taveta pond T. jipe males were also spawning at 21 cm. (though only about
120 gm. weight), and females of 15-21 cm. (about 28-120 gm.) were ripe. The
ovaries of ripe T. jipe from the Taveta pond were dark yellow and counts of ripe
eggs in the ovaries showed that 200-250 eggs are laid at a time by these small females.
T. jipe females are oral incubators; females were found with up to 167 eggs in the
mouth and others were seen carrying yolked young. In the Taveta pond females as
small as 15 cm. long were found brooding eggs. T. girigan is also a mouth brooder ;
up to 60 eggs were found in the mouth of one female. No males of either species
were found carrying eggs.

Although no Tilapia nests were seen on the bottom of the Taveta pond, some were
examined when the Korogwe pond stocked with Lake Jipe Tilapia was emptied.
These nests (PI. 17A) were identical in appearance with those made by T. pangani.
Each nest consisted of two or three series of small pits around a larger central pit, the
whole excavation being about two and-a-quarter feet across. It is not yet known
for certain which of the two Lake Jipe Tilapias made this nest, but they were pro-
bably made by T. jipe, and similar nests have been reported from the Taveta ponds
in which T. jipe predominated. It was surprising to find similar nests made by
T. jipe (or T. girigan) and T. pangani and this observation needs confirmation by a
more reliable correlation between nest and species responsible for it. For comparison
with these nests Plate 17B shows the simple circular hollows, about twelve inches in
diameter, made by T. esculenta in one of the Korogwe ponds.

The scales of 25 pond-living Tilapia showed that in most (88%) cases the immature
fishes have uniform scales whereas the scales of the mature show "checks " ; these
checks were near the edge of the scales in fishes which had recently spawned. In
males the checks appear as a band of circuli spaced more closely than on the rest of
the scale. In the females the checks are clear spaces, as though several circuli had
been eaten away ; it is thought that these represented spawning marks. Several of
the females grown from fry in the Taveta pond showed three (and one showed four)
such " spawning marks " after about fifteen months in the pond. Such checks
should prove useful for comparison of the growth of mature Tilapia under different
conditions.

The two species of Tilapia had not been distinguished when the samples of stomachs

NEW SPECIES OF TILAPIA FROM EAST AFRICA 363

and intestines were collected from Tilapia from the Lake, and data from pond fish
were influenced by the food available in the pond. Hence the data are not sufficient
to show any difference in feeding habits between the two species. Among the preserved
sample of Tilapia from Lake Jipe, however, ten T. girigan contained food, mostly
chewed water weed (Najas sp.) with a good supply of epiphytic diatoms and other
algae, or vegetable debris and algae from the bottom of the lake. The two T. jipe
preserved from Lake Jipe did not provide any information on the food of this species
under natural conditions ; the dense felt of pharyngeal teeth in this species suggests
that these fish feed on very small particles, probably algae, possibly phytoplankton.
The guts of ten other Tilapia from the lake, probably mostly T. girigan, contained
the same kinds of vegetable matter as the preserved T. girigan. In several cases
both plant material and bottom debris occurred in different parts of the alimentary
tract of individual fish showing that the individual Tilapia varies its feeding habit.
The pond fish, both species, had been eating bottom debris or mud with very little
organic matter. The size of the particles of bottom debris, etc., in the guts of the pond
fish did, however, reflect the degree of coarseness of the pharyngeal teeth, the stomach
and intestines of T. girigan containing larger particles than did those of T. jipe,
which has much more dense pharyngeal teeth. As the rectum of Lake Jipe Tilapia
in many cases contained food which was originally very different from that in the
stomach it was difficult to be certain how much of the food eaten was used. It was
clear, however, that diatoms were digested, that some plant cell contents were
extracted, but that much of the plant material eaten and the bluegreen and green
algae (including filamentous greens) pass through these Tilapia unused.

Most of the Tilapia examined from Lake Jipe carried large numbers of nematodes
in the pericardium. Some of the pond fish were similarly parasitised ; the apparently
more numerous parasites in the lake fish may be related to the more numerous
aquatic birds on the lake.

B. T. pangani

A sample of 39 Tilapia originally caught in traps in the Pangani River and living
in one of the Korogwe breeding ponds was examined. The 27 males ranged from
21 to 32 cm. long with a pronounced mode at 30 cm. and the females from 22 to 32 cm.
with a mode at 24 cm. (Text-fig. 3c). The size of these fish and the sex ratio in this
sample may perhaps have been due to selective action by the trap in the river. This
pond also contained numerous fry of 1-10 cm. long and the nest identical in appear-
ance with that shown in Plate 17A. These adults had been in the ponds for two
years ; it was said that they had not grown much during this time and that batches of
young were found every six weeks throughout the year. The pond contained little
food for these fish.

c. T. mossambica korogwe

Nothing is known of the habits of this small Tilapia found together with T. varia-
bilis and T. pangani in various ponds at Korogwe. It seems most probable that it
gained access to the ponds with the water supply from the Pangani River. The

ZOOL. II, 12. 22

364 NEW SPECIES OF TILAPIA FROM EAST AFRICA

T. m. korogwe were very docile, lying quietly on the bottom of the pond when the
pond was emptied and in contrast with T. variabilis which leapt about the mud
very actively in similar circumstances. Males as small as 16 cm. long were ripe
though they did not have a marked breeding dress. Intestines contained insect
remains and algal filaments. In spite of little food in the ponds these fish were in
fairly good condition which suggests that the breeding size of this species in the
natural environment may be similar to that in these ponds. The largest specimen
seen was 18 cm. long. The only female seen (to eight males) was 12 cm. long and the
ovary was already ripening.

D. Lake Chala Tilapia, T. hunteri Giinther

Lake Chala, an isolated lake on the foothills of Kilimanjaro in the Pangani drainage
area, was visited on 15th January, 195 1. The Tilapia from this lake, T. hunteri
Giinther, are endemic and of particular interest since there is no apparent inflow or
outflow to this deep, clear lake. Only immature Tilapia were caught. T. hunteri
has been described as having four anal spines; of the twenty-one caught, however,
only seven had four anal spines, the other fourteen had three. The shores of Lake
Chala are rocky and shelve steeply into deep water. The immature Tilapia were in
small shoals, each shoal of similar sized Tilapia, feeding on algae and debris off the
bottom between the rocks near the shore ; large numbers of crabs [Potamon (Potamon-
autes) platycentron (Hilgendorf)] were also living among these rocks. T. hunteri
may be distinguished from T. pangani, T. girigan, and T. jipe by the long shallow
body (body depth 36 %of standard length in 21 cm. fish), long narrow caudal
peduncle (length /depth 1*35 in 21 cm. fish) and a small narrow toothed area on the
pharyngeal bone. T. hunteri appears to be endemic to Lake Chala ; its relationship
with other Tilapia is not yet known.

IV. LAKE VICTORIA SPECIES OF TILAPIA GROWN IN PONDS

T. esculenta Graham and T. variabilis Boulenger have been introduced into the
experimental ponds at Korogwe. These species are endemic to Lakes Victoria and
Kyoga and do not occur naturally in the Pangani System, but data concerning their
growth in these ponds are given here, as in several respects they behaved rather
differently from the Pangani system Tilapia grown in the ponds. The fry (less than
20 mm. long) of both species were taken from the mouths of brooding female fish
caught near Mwanza at the south end of Lake Victoria and were introduced into a
small pond at Korogwe on 16th June, 1950. These fish were examined on 21st
January, 1951, six and three-quarter months later, and the following observations
were made.

T. esculenta Graham

There were 9 males and 7 females surviving. The males had grown to 16-17 cm -
and were ripe, the females to 17-19 cm. and had spawned (Text-fig. 5a). The
bottom of the pond had 46 nests, although only 16 breeding fish were present.

NEW SPECIES OF TILAPIA FROM EAST AFRICA 365

Each " nest " was a simple circular hollow in the bottom mud, about twelve inches
across and a few inches deep (PL 17B) ; these nests were in about three feet of water.

Numerous T. esculenta fry up to 7 cm. total length were seen. It was said that
this pond had been emptied and fry observed early in December, only six months
after the original fry were put into the pond.

These Tilapia had very little nutritious food. Stomach contents showed
that they had been feeding on bottom mud and diatoms. Many of the diatoms in the
stomach were already empty or half empty and some diatoms in the rectum still
contained some contents, possibly because of the amount of mud passing through
the gut at the same time which might impede digestion.

In Lake Victoria male and female T. esculenta are generally at least 20-25 cm.
long before they start to spawn. These Tilapia taken from the mouths of " normal
sized " Lake Victoria Tilapia were breeding when 16-19 cm - l° n g> an d six and three-
quarter months old.

T. variabilis Boulenger

There were 12 males and 13 females surviving. The males and females were
growing at the same rate, fish of both sexes having grown to 13-19 cm. (with a mode
at 16 cm), in six and three-quarter months (Text-fig. 5b). A number of the fish
were ripe, but only one nest was seen. This was a simple circular hollow like the
T. esculenta nests and like many of the T. variabilis nests seen in Lake Kyoga in
Uganda.

Numerous fry 1-6 cm. long were present. It was said that no fry had been
seen when the pond was emptied in December, 1950.

These T. variabilis were noticeably more active than the other Tilapia species,
leaping about in the mud when the pond was emptied.

They had not grown quite so fast as the T. esculenta. Males and females showed
the same rate of growth which bears out observations on Lake Victoria. Both
sexes were breeding at 16 cm. long and when less then seven months old under these
conditions, though in Lake Victoria from whence the fry came they would probably
not start to breed until about 20 cm. long.

V DISCUSSION

The few odd specimens from the Pangani which do not fit into any of the
new species described here suggest that there are further Tilapias in the Pangani
system. Further field work is essential for sorting out these species and finding out
how the various species keep distinct. Such field work needs to be done immedia-
tely as the introduction of species foreign to the Pangani system into ponds, and the
probability that some will escape into the river, makes the task of unravelling species
increasingly difficult and the results of less value in throwing light on the origin,
evolution and distribution of these Tilapias.

It was expected that differences in breeding colours and nests would help to keep
the species distinct, and it is surprising to find breeding colours so similar in the two

3 66

NEW SPECIES OF TILAPIA FROM EAST AFRICA

species of Lake Jipe, and that the nests of these and of T. pangani should be alike.
Further observations are required in the light of the definitions of the species here
presented.

Among the T. girigan and T. jipe grown in the Taveta ponds the males were
considerably larger than the females of the same age. Either the males grow faster

A A

6
>

111

£3

1 \

• •

I I I

Si
7

6-1
>

UJ
£34

t T . V *' J

5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

TOTAL BODY LENGTH - CENT IM ET RE S

Text-fig. 5. Length frequencies of Lake Victoria Tilapias grown in fishponds, six and
three-quarter months after the ponds were stocked with fry less than 20 mm. long.
Males , Females . (A) T. esculenta. (B) T. variabilis.

than the females or they continue to grow after growth in the females has slowed
down. It would be interesting to know how much this is an inherent character of
these species and how much it may be influenced by the very unequal sex ratio, the
more numerous males having more energy available for growth than the females
who may be constantly engaged in reproduction. Among the T. variabilis, males
and females grew at approximately the same rate, but there were approximately

NEW SPECIES OF TILAPIA FROM EAST AFRICA 367

equal numbers of males and females. Among T. variabilis and T. esculenta in Lake
Victoria males and females are of comparable sizes and appear to grow at the same
rate. The males of T. mossambica Peters are larger than the associated females, and
Baerends and Baerends van Roon (1950) have shown that in this species 1 the size of
the male is important in allowing him to establish and keep a spawning territory.
Further observational and experimental work is needed to elucidate why sexual
dimorphism and size differences exist in some species of Tilapia and not in others.

The snout of mature male T. girigan, T. pangani, and to a lesser extent T. jipe
becomes elongated and the fish develops a very concave upper profile, as in T.
mossambica Peters. The few large females also tended to develop the " male "
profile, though to a lesser extent. Thus it seems that the elongation of the snout is
partly a sexual character and partly a growth character, the snout showing positive
allometric growth in relation to the growth of the fish.

This study has shown a variation in the number of anal spines in three species
of Tilapia, T. jipe, T. m. korogwe and T. hunteri Giinther ; this character was
previously considered to be of considerable stability and specific significance in this
genus. Trewavas (1937) described fossil Tilapia with four anal spines from Pleisto-
cene deposits at Rawe near Lake Victoria and concluded that they belonged to
T. nigra Giinther. Greenwood (195 16) discussed fossil Tilapia with four anal
spines from Miocene deposits on Rusinga Island, Lake Victoria, and considered these
were most closely related to T. mossambica Peters, especially specimens from the
Tana system (types of Chromis spilurus Giinther 1894). Summarizing the geological
evidence in connection with the distribution of Haplochromis species in East Africa,
Greenwood (1951a) concluded that the eastward flowing rivers probably provided
retreats for fish from Lake Karunga during the Miocene drying-up period. Lake
Victoria now lies in part of the area formerly covered by Lake Karunga and it seems
likely that the Tilapia found today in the Pangani system may have had a common
origin with the four spined forms now fossilized near and in lake Victoria. Dr. E.
Trewavas of the British Museum (Natural History) is at present studying the Tilapia
of the Tana and other eastward-flowing rivers of East Africa. Her results are
awaited with great interest and it is hoped that they will illuminate the relationships
and probable lines of evolution of the species described here.

SUMMARY

Three new species and one new subspecies of Tilapia are described from the Pan-
gani system, two species from Lake Jipe, T. jipe and T. girigan, and from the Pangani
River a new species, T. pangani, and subspecies T. mossambica korogwe. T. girigan
and T. pangani appear to be members of the T. mossambica complex. A few speci-
mens do not fit into any of these species which suggests that there are further Tilapias
in the Pangani. Further field work is necessary to see how the species keep distinct.

T. jipe has a vertically striped caudal fin, previously regarded as a specific "spot
test " for T. nilotica Linn, in East Africa. The number of anal fin spines, a character

1 T. natalensis = T, mossambica Peters (Trewavas 1937).

368 NEW SPECIES OF TILAPIA FROM EAST AFRICA

generally of specific significance among Tilapia, was found to vary in T. jipe, T. m.
korogwe and T. hunteri Gunther.

Length frequencies of lake and pond fish showed that in both T. girigan and
T. jipe : (a) the males were considerably larger than the females ; (b) the females
started to spawn at a smaller size than the males ; (c) males were much more
numerous than females in the breeding population. " Nests " of Lake Jipe Tilapia
and T. pangani are described and compared with the nests of T. esculenta Graham
in neighbouring ponds.

Observations are given on the growth of T. esculenta Graham and T. variabilis
Boulenger reared in ponds from fry from Lake Victoria. In both these species
males and females showed the same rates of growth and started to breed at the same
size. Both T. esculenta and T. variabilis in these ponds were breeding when 16 cm.
long and less than seven months old, though in Lake Victoria, from whence the fry
came, they would probably not start to breed until about 20 cm. long.

ACKNOWLEDGEMENTS

I would like to thank Colonel E. S. Grogan, and Major R. E. Gould, Fish Culturist
in charge of the Korogwe Fish Farm, for their invaluable help while collecting the
data on which this paper is based. I am very grateful to the Director of the East
African Fisheries Research Organization and other members of the staff who have
helped with stimulating discussion and criticism. To Dr. E. Trewavas of the British
Museum (Natural History) my especial thanks are given for assistance and encourage-
ment with the systematic side of the investigation. Plates 13-16 are the work of
the staff of the photographic section of the British Museum (Natural History).

REFERENCES

Baerends, G. P., and Baerends van Roon, J. M. 1950. An Introduction to the study of the

ethology of Cichlid fishes. Behaviour. Supplement 1, pp. 1-242. Leiden.
Boulenger, G. A. 1915. Catalogue of the Freshwater Fishes of Africa. Vol. 3. (London).
Greenwood, P. H. 1951. Evolution of the African Cichlid fishes : The Haplochromis

species flock in Lake Victoria. Nature. London, 167 : 19.
■ 195 1. Fish remains from Miocene deposits of Rusinga Island and Kavirondo Province,

Kenya. Ann. Mag. nat. Hist. (12) 4 : 1192.
Pfeffer, G. 1896. Die Thierwelt Ost-Afrikas. Lief. V Fische (p. 10).

Playfair, R. L., and Gunther, A. C. L. G. 1866. The Fishes of Zanzibar. London (p. in).
Trewavas, E. 1937. Fossil Cichlid fishes of Dr. L. S. B. Leakey's Expedition to Kenya in

1934-5. Ann. Mag. nat. Hist. (10) 19 : 381.

PLATE 13

Tilapia jipe, a male of 25 cm. total length from one of the Taveta fishponds.

PLATE 14
Tilapia girigan, a male of 23 cm. total length from Lake Jipe.

PLATE 15
Tilapia pangani, a male of 28 cm. total length from one of the Korogwe fishponds.

PLATE 16

Lower pharyngeal bones of Tilapia species. a. T. jipe (fish 23 cm. total length), b. T.
girigan (fish 23 cm.), c. T. pangani (fish 23 cm.), d. T, mossambica korogwe (fish 20 cm.).
e. T. mossambica (fish 20 cm.), f. T. nilotica (fish 22 cm.)

Bull. B.M. [N.H.) Zool. II, 12

PLATE 16

TIL API A

PLATE 17

a. Nest of Lake Jipe Tilapia in one of the Korogwe ponds. The whole
excavation is 2jft. across (the ruler is 15 in. long).

b. Nests of Tilapia esculenta from Lake Victoria in one of the Korogwe ponds.
Each excavation is about 12 in. across.

hull. B.M. (N.H.) Zool. II, 12

PLATE 17

TIL API A

TIL API A ESCULENT A

PRESENTED

2 JUL 1955

PRINTED IN GREAT BRITAIN BY
ADLARD AND SON, LIMITED,

BARTHOLOMEW PRESS. DORKING

2 SEP 1955

THE MORPHOLOGY OF THE HEAD
OF THE HAWFINCH

(COCCOTHRAUSTES COCCOTHRAUSTES)

R. W. SIMS

BULLETIN OF
THE BRITISH MUSEUM (NATURAL HISTORY)
ZOOLOGY Vol. 2 No. 13

LONDON: 1955

THE MORPHOLOGY OF THE HEAD OF

THE HAWFINCH

(COCCOTHRAUSTES COCCOTHRAUSTES),

WITH SPECIAL REFERENCE TO THE
MYOLOGY OF THE JAW

REGINALD WILLIAM SIMS

Vvi

Pp.369 — 393 ; 10 Text-figures

BULLETIN OF

THE BRITISH MUSEUM (NATURAL HISTORY)

ZOOLOGY Vol. 2 No. 13

LONDON: 1955

THE BULLETIN OF THE BRITISH MUSEUM
(NATURAL HISTORY), instituted in 1949, is
issued in five series corresponding to the Departments
of the Museum, and an Historical Series.

Parts appear at irregular intervals as they become
ready. Volumes will contain about three or four
hundred pages, and will not necessarily be compiled
within one calendar year.

This paper is Vol. 2, No. 13 of the Zoological series.

PRINTED BY ORDER OF THE TRUSTEES OF
THE BRITISH MUSEUM

Issued August, 1955 Price Eight Shillings

THE MORPHOLOGY OF THE HEAD OF
THE HAWFINCH

(COCCOTHRAUSTES COCCOTHRAUSTES),

WITH SPECIAL REFERENCE TO THE
MYOLOGY OF THE JAW

By R. W. SIMS

CONTENTS

Page
Introduction . . . . . ... . . . . 371

Materials .......... . 372

Acknowledgments . . . . . . . . . .372

Historical Note . . . . . . . . . .372

Description ........... 373

1. Rhamphotheca ......... 373

2. Skull ........... 374

3. Myology of the Jaw . . . . . . . 376

Comparative Note .......... 385

Discussion ........... 389

Conclusions ........... 390

Appendix ........... 391

References ........... 392

SYNOPSIS

The ability of the Hawfinch, Coccothraustes coccothraustes , to crack open the stones of cherries,
damsons and olives in order to feed on the kernels is generally well known, but the modifications
which are present in the structure of the head and enable this comparatively small bird to perform
such a feat have apparently received little attention. The method employed by C. coccothraustes
to crack open the fruit stones is noteworthy because each stone is held between the mandibles
and is broken solely by the force applied by the jaw muscles, and not by any " artifice " such as
that employed by the thrush to smash snail shells, or the woodpecker to split almond stones.
The behaviour of the bird when feeding on fruit stone kernels appears to follow a definite pattern.
The bird usually selects fallen fruit apparently discarding the soft parts. Observations and photo-
graphs show that a cherry stone is positioned in the mouth by the combined movements of
the head and tongue until it is held lengthways between the mandibles at the back of the horny
palate of the mouth with the suture of the stone lying in the median sagittal plane of the
head. A quick snap of the jaw and the shell is neatly cracked along the suture or " seam ".
The halves of the shell are rejected and the kernel is swallowed whole, usually without being
crushed. Experiments on breaking open cherry and olive stones (given in detail in the appendix)
show that pressures in the region of about 100 lb. are required to perform this feat 1

zool. 2,13. 21

372 THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH

The purpose of this paper is to show how the head of C. coccothraustes is adapted to apply
and withstand such forces by describing some aspects of the morphology of the head, namely, of
the horny bill, or rhamphotheca, the skull and the jaw muscles.

MATERIALS

This study is based on specimens of C. coccothraustes collected in Great Britain.
The material consists of two adults and two juvenile male specimens preserved in
spirit in the National Collection, one head, sex unknown, supplied by Mr. G. R.
Mountfort, and several skeletons in the National Collection. In most of the osteo-
logical material the rhamphothecae, which can only be removed with difficulty,
were intact. The heads of two spirit specimens were dissected completely during
the course of the investigation. The head of a spirit specimen of a Brambling,
Fringilla montifringilla, was also dissected, and reference was made to the series of
skulls of the Chaffinch, Fringilla coelebs, in the National Collection.

ACKNOWLEDGMENTS

I am indebted to Mr. Guy Mountfort for suggesting this investigation. The work
was originally undertaken in order to answer some of the points raised by him in the
preparation of his monograph of the Hawfinch. The scope of this study was subse-
quently enlarged. I am obliged to Mr. H. L. Cox and his colleagues of the National
Physical Laboratory, and Mr. D. Welbourn of Cambridge, for providing me with the
interesting data which form the appendix to this paper ; and to Dr. J. Wahrman
of the Hebrew University, Jerusalem, for a supply of fresh olives for the crushing
tests. I also wish to thank my colleagues in the British Museum (Natural History)
for reading the MS. and for providing many helpful criticisms.

HISTORICAL NOTE

References to the general anatomy of the head of C. coccothraustes are remarkably
few and are confined mainly to the consideration of characters of the skull as an aid
to classification, or less frequently to its architecture. The skull with its high degree
of ossification particularly impressed W. K. Parker (1879) by what he called " its
ridgy strength " when he described some aspects of the anatomy of the skull in his
systematic work on the structure of the palates of passerine birds. At the beginning
of the section on C. coccothraustes he pointed out that most of the differences between
the skull of this species and those of other members of the Fringillidae were of little
phylogenetic importance because the modifications in the skull of C. coccothraustes
were for " mechanical purposes". Nevertheless, Bowdler Sharpe (1888) used many
of these modifications " for convenience " to separate what he termed the Cocco-
thraustinae from the buntings and true finches.

The peculiarities of the horny bill or rhamphotheca were noted very briefly by
Pycraft (1905), and his observations were occasionally referred to by subsequent
workers. Nothing new was recorded on the horny bill until Sushkin (1925) described
similar modifications in the bill of the Evening Grosbeak, Hesperiphona vespertina.

THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH

373

He concluded that they were essentially the same although purely adaptive charac-
teristics. He agreed with W. K. Parker that most of the peculiarities of the skull
were adaptive features, but believed that only one character of the bony palate was
of systematic significance, and on that he regarded the Evening Grosbeak as a member
of a sub-family containing the Old World hawfinches.

Later references are confined almost entirely to remarks in papers dealing with the
mechanical structure of bird skulls, for example those by N. G. von Lebedinsky
(1921) and von Kripp (1933a). The latter author showed that the skull was struc-
turally stronger than that of a harrier.

The literature appears to contain few references to the myology of C. coccothraustes,
so an account of the musculature of the jaw is given in this present paper. The
nomenclature adopted is that proposed by Lakjer (1926) in his comparative work on
the Sauropsida.

DESCRIPTION

1. Rhamphotheca

The rhamphotheca of birds is derived from the malpighian layer of the epidermis.
In most birds it is seldom more than a thin sheath which is readily detached from a
dried skull, but in C. coccothraustes parts of it inside the mouth are greatly thickened
and enlarged. In the palatal region (Text-fig. za) it forms a longitudinally striated
thick pad partly divided by a longitudinal median depression. This thickened area
extends from the anterior border of the palatine bones over the posterior third of the
premaxillae. The anterior two-thirds of the premaxillae are covered by the
remainder of the horny palate which is strengthened by one median and a pair of
lateral ridges. The grooves formed between the lateral ridges and the edges of the
upper mandible accommodate the edges of the lower mandible. On the lower jaw
(Text-fig. ib) the horny sheath is thickened to form two large bosses which lie

Tongue

Fig. i. Coccothraustes coccothraustes. The oral surfaces of the rhamphotheca
showing the striated pads, a, Upper jaw ; b, Lower jaw with the anterior of the
tongue in situ.

ZOOL. 2, 13. 2I§

374 THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH

immediately over the posterior parts of the dentary bones which are themselves
specially thickened in this area. The bosses are striated, but the striations being
curved more or less follow the contours. The depression between the two bosses
accommodates the tongue. The tip of the tongue is broadly cuneiform in shape
and lies flush with the anterior surfaces of the lateral bosses.

2. Skull

The principle modifications of the skull which appear to be connected with the
ability of C. coccothraustes to exert a great force with its jaw muscles can be
summarised as follows :

In the first place the whole skull is strengthened by a greater ossification than is
usually found in most fringilline species. An indication of the extent of ossification
as determined by weight, is given in a later section.

Von Kripp (1933a) described how the architecture of the skull is suited to withstand
stress and strain. Viewed laterally the outline of the skull approximates to one
horn of a crescent, the dorsal profile being convex and the central one concave
(Text-fig. 2a). The concave ventral surface is strengthened by a system of struts
formed by a massive pyterygoid-quadrate-zygoma system. The anterior part of
the head is further strengthened by the heavily ossified nasal and interorbital septa.
These elements play an important role in the rigidity of the skull for their function
is analogous to that of the vertical component of an I-shaped girder.

In the skulls of many birds the whole of the upper mandible is hinged to the
cranium, but in C. coccothraustes it is rigidly fixed. The suture between the frontal
and nasal bones, where the hinge is usually located, is obliterated ; the nasal and
the interorbital septa form what is functionally a continuous vertical wall ; and the
palatines are ankylosed with the rostrum and vomer which form most of the base
of that wall.

The fixity of the upper mandible also may be associated with the nature of the
relationship between the cranial head of the quadrate and the surrounding structures
in the otic region of the skull. The quadrate is joined to the palatine by the ptery-
goid and to the upper mandible by the zygoma, and owing to the immobility of the
upper mandible the quadrate is held firmly by these bones. The absence of move-
ment together with the great crushing strain imposed on the quadrate by the lower
jaw has resulted in the squamosal and opisthotic facets in the otic region extending
partly around and in close contact with the small squamosal and opisthotic heads
of the quadrate. The extension of the area of contact between the quadrate and the
cranium is functionally possible only because the quadrate is held firmly by the
pterygoid and the zygoma. The increased area of contact provides a strong
base for the quadrate which in its capacity as the fulcrum of the lower jaw experi-
ences great pressures when hard food, such as fruit stones, are cracked.

The great pressures have also influenced the nature of the quadrate which is very
massive. Moreover, a powerful muscle, M. Quadrato-mandibularis, originates from
its orbital process, and the inclusion of the muscle as an important member of the

THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH

375

Adductor Group has no doubt contributed towards the massive development of this
bone.

The post-orbital process of the squamosal is unusually large and this feature may-
be correlated with the powerful muscle, the pars medius of the M. mandibulae
externus, which has a tendinous origin on it. The pars medius is inserted into the

post. orb. proc.

proc. end.

int.

pars post.

for. md.

Fig. 2. Coccothraustes coccothrausles. a, cranium and upper jaw, lateral view
from the right side, b, Lower jaw. Ant. pal. proc, anterior process of the palatine ;
for. md., foramen mandibularis ; pars, ant., anterior part of the lower jaw (dentary) ;
pars, post., posterior part of the lower jaw ; post. orb. proc, post-orbital process of the
squamosal ; post. pal. proc, posterior process of the palatine ; proc end., processus
coronoideus ; proc md. int., processus mandibulae internus (articular) ; pt., pterygoid ;
quad., quadrate; zy., zygoma (quadrato-jugal and jugal).

lateral surface of the processus coronoideus, but the modifications associated with
this will be discussed later.

Dislocation of the lower jaw from the quadrate during the contraction of the power-
ful jaw muscles is prevented by a ligament which extends from the zygoma to the
posterior surface of the processus mandibulae internus, passing over the posterior

376 THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH

surface of the distal end of the quadrate. This ligament has two centres of ossifica-
tion ; one forms a large sesamoid which lies close to, and posterior to, the articula-
tion of the jaw ; the other is smaller and lies more laterally near the attachment
of the ligament to the zygoma.

The lower jaw, like the remainder of the skull, is massive and ossified to the extent
that the foramen mandibularis, usually a large opening, is reduced to a size that
permits only the passage of a branch of the trigeminal nerve (Text fig. 20).

The anterior part of the lower jaw is strengthened by an inward expansion and
union of the paired dentaries, the long symphysis forming a bony shelf. Postero-
mesially the dentaries are greatly thickened to serve as strong foundations for the
heavy bosses of the rhamphotheca.

3. Myology of the Jaw

The muscles involved in the movement of the jaws of birds form four functional
groups :

(a) The Adductor Group.

(b) The Constrictor Group.

(d) The Retractor Group.

The Adductor Group raises the lower jaw, the Constrictor Group depresses it ;
the Protractor Group raises the upper mandible and the Retractor Group lowers it.
In C. coccothraustes the upper mandible is not hinged on the cranium, as already
mentioned, and the muscles of the Protractor Group which persist are functionless,
while the muscles of the Retractor Group assist in the elevation of the lower jaw
and are dealt with here as members of the Adductor Group.

The Constrictor Group and the Protractor Group in C. coccothraustes each
consists of only one paired muscle and therefore neither in sensu stricto can be
considered as a "group". Nevertheless, the term has been used here for convenience
in both cases.

I The Adductor Group

The Adductor Group, which raises the lower jaw, is the largest of the three groups.
This assemblage in C. coccothraustes may be further divided into two clearly marked
functional units, one acting from the top and back of the cranium and the other
acting from the orbital walls and their vicinity. The muscles of the former are
inserted in the ramus of the lower jaw near the superior (dorsal) and the posterior
margin of the coronoid elevation. Functionally, this is the more important unit
and consists of the Ms. Adductor mandibulae and the M. Quadrato-mandibuteris.
The other muscles of the Adductor Group, the M. Pterygoideus and the M. Ethmo-
mandibularis, are inserted near the inferior (ventral) margin of the ramus of the
lower jaw.

the morphology of the head of the hawfinch 377

Description of the Adductor Group

M. ADUCTOR MANDIBULAE EXTERNUS (Text-fig. 3)

(1) Pars superficialis

The rostral part is fan-like and arises from the cranium over most of the anterior
half of the frontal. The area of attachment is confined anteriorally and mesially
by the frontal crest, and laterally by the flattened rim of the orbit. The fibres
pass forwards and downwards between the post-orbital process of the squamosal

post. orb. proc.

Dp. md.

Fig. 3. Coccothraustes coccothraustes. Diagrammatic view of the right side of
the head showing the external jaw muscles, p.m., p.p., p.s., partes medius, profundus,
superficialis of the M. Adductor mandibulae externus; Dp. md., M. Depressor
mandibular.

and the rim of the orbit. The fibres converge becoming more tendinous and are
inserted at the apex of the processus coronoideus of the lower jaw. The main part
extends over the posterior half of the frontal ; it is bounded by the rostral part in
front and the transverse post-frontal crest behind, and extends inwards to the
meso-frontal crest. The converging fibres are directed more forwards than those
of the previous part and pass over its aponeurosis and the post-orbital process
of the squamosal. The head is inserted in the pars posterior of the lower jaw
laterally to the head of the rostral part at the apex of the processus coronoideus.
The caudal part arises from that area of the cranium bounded by the transverse
post-frontal crest in front and the transverse crest on the posterior part of the
squamosal and the parietal. (Mesially the two prominences converge and form a
triangular area of attachment). The fibres passing more forwards than downwards
become more tendinous as they converge; they are inserted somewhat laterally in
the posterior margin of the processus coronoideus. This part is similar in size to

378 THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH

the other two but it differs in that the converging fibres pass under both the distal
part of the post-orbital process of the squamosal and the ascending fibres of the
pars medius (see below) .

(2) Pars profunda

The pars profunda arises from the cranium on the anterior surface of the supra-
meatal ridge of the squamosal and in the concavity in the otic region between the
cranium and the quadrate, also to the neck of the quadrate. This muscle is very
tendinous throughout although it is inclined to be more fleshy in the part occupying
the concavity in the otic region. It is inserted in the ramus of the lower jaw ventrally
to the site of insertion of the pars superficialis, that is, about half-way along the outer
edge of the posterior margin of the processus coronoideus. Some of the fibres and
the dense fascia surrounding the muscle coalesce with those of the pars superficialis.

(3) Pars medius

The pars medius is a fan-like muscle with a tendinous origin on the ventral surface
of the post-orbital process of the squamosal. The fibres pass downwards and become
more fleshy as they diverge. The muscle is inserted over most of the lateral surface
of the processus coronoideus of the lower jaw.

M. ADDUCTOR MANDIBULAE POSTERIOR (Text-fig. 4)

The M. adductor mandibulae posterior arises from the posterior wall of the orbit,
namely, from the alisphenoid outside the latero-ventral margin of the fossa in the
orbital wall. Mesially there is a thin bony partition which separates the muscle
from the M. Pseudotemporalis (see below), and laterally another thin bony promi-
nence extends down from the rim of the orbit and separates the muscle from the
descending fibres of the pars superficialis. The muscle is short and very tendinous.
It is inserted in the mesial surface of the processus coronoideus near the upper
half of the posterior margin. There are a few tough ligamentous strands extend-
ing the whole length of the muscle and these are inserted in the small dorsally
directed conical process of the mesial surface of the processus coronoideus near the
posterior margin.

M. ADDUCTOR MANDIBULAE INTERNUS (M. PSEUDOTEMPORALIS)

(Text-fig. 4)

The M. Pseudotemporalis arises on the posterior wall of the orbit from the alis-
phenoid below the fossa in the orbital wall. The site of attachment is separated
laterally from the M. Adductor mandibulae posterior by a thin bony partition (see
below), and mesially from the optic foramen by a similar partition. The muscle is
tendinous throughout and it is attached to the ramus of the lower jaw in the mesial
surface of the processus coronoideus. The site of attachment is a little below the
antero-dorsal margin of the ramus anterior to the M. Adductor mandibulae posterior,

THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH 379
Pt. d. m.

Fig. 4. Coccothraustes coccothraustes. a, Diagrammatic view of the right side of
the head, eye and eye muscles removed to expose the jaw muscles in the orbit, b, Dia-
grammatic dorsal view of the right orbit. A .m.p., M. Adductor mandibulae posterior ;
E. md., M. Ethmo-mandibularis ; Pst., M. Pseudotemporalis (M. Adductor mandibulae
internus) ; Pt.d.l., M. Pterygoideus dorsalis lateralis; Pt.d.m., M. Pterygoideus
dorsalis medialis ; Qmd., M. Quadrato-mandibularis ; //, Optic nerve.

M. QUADRATO-MANDIBULARIS (Text-fig. 4)

The M. Quadrato-mandibularis forms a thin sheet of muscle, almost devoid of
tendinous tissue, underlying the Ms. Pseudotemporalis and Adductor mandibulae
posterior. It extends from the quadrate to the mesial surface of the processus
coronoideus. The muscle arises from both surfaces of the orbital process and as
far as the neck of the quadrate where part of the pars profunda of the M. Adductor

380 THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH

mandibulae externus originates. The muscle is inserted obliquely in the ramus
of the lower jaw along a line tending to pass from below the site of insertion of the
M. Pseudotemporalis towards the point of articulation with the quadrate.

M. ETHMO-MANDIBULARIS (Text-fig. 5)

The M. Ethmo-mandibularis arises from the anterior part of the interorbital
septum above the palatine and partly from the anterior wall of the orbit, that is,
from the mesethmoid. Some of the dorsal fibres pass horizontally across the
anterior angle of the orbit and assist in the support of the main fibres which pass
obliquely downwards and outwards back to that part of the lower jaw behind the
gape. The fibres become tendinous a little before being inserted into the mesial
surface of the ramus of the lower jaw above the ventral margin.

M. PTERYGOIDEUS (Text-fig. 5)

M. PTERYGOIDEUS dorsalis

M. Pterygoideus dorsalis medialis

The M. Pterygoideus dorsalis medialis arises from the interorbital septum at the
antero-dorsal angle of the orbital wall above the attachment of the posterior part
of the M. Ethmo-mandibularis. The muscle has numerous tendinous fibres near
its site of origin, and these increase in number until the muscle is completely tendinous
a little before the site of insertion. The muscle passes backwards and downwards
between the ventral surface of the orbital process of the quadrate and the dorsal
surface of the pterygoid (beneath the M. Protractor pterygoidei, see below). The
muscle is attached by a very strong tendon to the ramus of the lower jaw at the head
of the processus mandibulae internus. The tendon lies over the head of the inflexion
and is inserted into the upper part of the posterior surface.

M. Pterygoideus dorsalis lateralis

The M. Pterygoideus dorsalis lateralis arises principally from the palatine and from
the flattened area at the union of the pterygoid and the palatine. The thick fleshy
fibres pass obliquely downwards and backwards and outwards to the site of insertion
which is towards the ventral margin of the mesial surface of the ramus of the lower
jaw, posterior to the M. Ethmo-mandibularis. The only tendinous tissue in the
muscle is in the nature of a thin sheet extending throughout the entire muscle.

M. Pterygoideus ventralis

M. Pterygoideus ventralis medialis

The M. Pterygoideus ventralis medialis arises from near the top of the mesial
surface of the processus mandibulae internus. It is a slender muscle and the fibres
pass forwards and upwards becoming more tendinous. The tendon is inserted in
the palatine along the outer (ventral) surface of the posterior process where it curves
as if to meet the other member of the pair.

THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH 381

Pt. d. m.

E.md.

v. m.

Pt. v. I

pars ant

ant. pal. proc

pars post.

Pt. v. I. a.

Pt. v. /. p.

proc. md. int.

Pt. v. m.

Pig. 5. Coccothraustes coccothraustes. a, Diagrammatic view of the right side of
the head dissected to expose the ventral jaw muscles, b, Diagrammatic ventral view
of the ventral jaw muscles. Pr.pt., M. Protractor pterygoidei ; Pt.v.l.a., anterior
portion of the M. Pterygoideus ventralis lateralis ; Pt.v.l.p., posterior portion of the
M. Ptergoideus ventralis lateralis ; Pt.v.m., M. Pterygoideus ventralis medialis.

382 THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH

M. Pterygoideus ventralis lateralis

Anterior Portion. The anterior portion of the M. Pterygoideus ventralis lateralis
is a thin fan-like muscle arising mainly from the ventral margin and partly from the
mesial surface of the ramus of the lower jaw. The site of attachment is mid- way
along the ventral margin of the pars posterior of the ramus. The fibres pass upwards
mesially to other muscles, then they become more tendinous and are inserted along
the outer member of the bifurcation at the distal end of the anterior palatine process.

Posterior Portion. The posterior portion of the M. Pterygoideus ventralis lateralis
is a stouter muscle than the anterior portion. It arises along most of the mesial
surface of the processus mandibulae internus of the lower jaw, that is, all of the
mesial surface except for the small area occupied by the M. Pterygoideus ventralis
medialis. It is also attached along a little of the ventral margin and mesial surface
of the main body of the ramus of the lower jaw. The fibres are markedly tendinous
and ultimately form a tough tendon which in inserted along the inner branch of the
bifurcation at the distal end of the anterior palatine process (the anterior portion is
attached to the outer branch).

Remarks on the Adductor Group

The relatively large area of attachment of the muscles of the Adductor Group
indicates that the total stress of the combined contraction is evenly distributed over
most of the cranium. The M. Adductor mandibulae externus arises over the lateral,
dorsal, and posterior surfaces of the cranium and the Ms. Adductor mandibulae
posterior and internus (M. Pseudotemporalis) arise over the anterior of the cranium,
that is, the posterior wall of the orbit. The Ms. Pterygoideus and Ethmo-mandibularis
are attached to the mesethmoid, palatine and the pte^goid and the stress of their
contraction is relayed to the antero-ventral surface of the cranium by the
interorbital septum.

Functionally, it is not strictly correct to include all the elements of the M. Ptery-
goideus in the Adductor Group of muscles, for when both the mesial members of the
M. Pterygoideus dorsalis and ventralis contract they seem to assist in the depression
of the lower jaw ; they draw the processus mandibulae internus of the articular
forwards, and so rotate the anterior of the lower jaw downwards through the same
angle (Text fig. 6) . However, on closer examination it becomes apparent that these
muscles and the posterior portion of the M. Pterygoideus ventralis lateralis hold the
articulatory surfaces of the lower jaw and the quadrate in contact. Thus they
strengthen the hinge mechanism and help prevent the dislocation of the jaw when
the powerful crushing muscles contract.

If the upper mandible of C. coccothraustes were hinged and the palatines were free
and formed part of this mechanism then those members of the Ms. Pterygoideus
dorsalis and ventralis which are inserted in the palatines would act as retractor
muscles. These would draw the palatines backwards so lowering the upper mandible
and holding it fast. However, it seems proper to include all the members of the
Ms. Pterygoideus dorsalis and ventralis in the Adductor Group for by their contrac-
tions they ensure the most efficient functioning of the main " adducting " muscles.

THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH 383

Not only is it valid to include these muscles in the same functional group because of
their necessary simultaneous contractions but because they are innervated by the
same branch of the trigeminal nerve.

<-M

«-M

rR .

1— . <D

CO" O

•+-> CO

•+J SH

■IH »H

bp p

00 CO

en ,_

4^.3

^ ft

J> rC

A £

^ H

<~ £

O >

■xJ

'-'

+» R

Sh " —>

**-• <D

•iH <-H

> O

■«

-+-> ■«

06* f

Q rn

s s

00 ,rt

&°

•- R

* ."S

lb M

>* o

3 '6

3 is

81
85

384 THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH

II The Constrictor " Group "

The depression of the lower jaw, as mentioned above, is effected by only one paired
muscle.

Description of the Constrictor " Group "

M. DEPRESSOR MANDIBULAE (Text-figs. 3 and 7)

The M. Depressor mandibulae arises from the greater part of the exoccipital,
especially the part forming the posterior wall of the bony meatus. The fibres of
the muscle pass downwards without converging, or becoming very tendinous, to
where they are inserted into the post-articular part of the mandible. Tendinous

Dp. md.

Fig. 7. Coccothraustes coccothraustes. Diagrammatic posterior view of the head
showing the external jaw muscles.

insertions are made in the angular behind the lateral process, ventrally in the posterior
of the angular, and in the posterior of the processus mandibulae internus. The fibres
form a flat muscle which completely covers the mandibular articulation from
behind and also forms a fleshy extension to the posterior wall of the meatus.

Remarks on the Constrictor "Group"

The M. Depressor mandibulae is a feeble muscle in comparison with the muscles of
the Adductor Group. Its action does not necessitate any rapid contraction, nor is
it required to exert any great force to depress the lower jaw. The chief function
of the muscle is to overcome the effect of the relaxed, but unextended, muscles of
the Adductor Group that would otherwise tend to oppose the depression of the
lower jaw.

THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH 385

III The Protractor " Group "

The upper mandible does not articulate on the cranium in C. coccothraustes but
although this mechanism is absent one functionless paired muscle, as already
mentioned, persists as a vestigial structure.

Description of the Protractor " Group "

M. PROTRACTOR PTERYGOIDEI (Text-fig. 5)

The M. Protractor pterygoidei is a slender muscle arising from the alisphenoid
in the posterior wall of the orbit a little below the optic foramen. The fibres pass
downwards and outwards beneath the orbital process of the quadrate ; they converge
and becoming tendinous are attached to a small dorsally situated spine-like process
near the quadratal end of the pterygoid.

Remarks on the Protractor " Group "

In species where the upper mandible is hinged on the cranium the function of the
M. Protractor pterygoidei is to draw the distal (quadratal) end of the pterygoid
upwards. The quadrate is attached to this and when the pterygoid moves upwards
the quadrate is rotated forward, thrusting the pterygoid forward at the same time.
This action pushes forward the palatine under the interorbital septum. The upper
mandible is hinged to the cranium on the dorsal side so that forward movement
on the ventral side communicated by the palatines has the effect of lifting the bill
on its hinge. The zygoma is also attached to the quadrate and the forward rotation
of the quadrate causes the zygoma to be moved forward, thus assisting in the eleva-
tion of the upper mandible since the zygoma is joined to the maxilla. Although
the muscle is functionless in C. coccothraustes is is interesting to note that its fibres
are not completely lacking in tendinous strands and generally the muscle closely
resembles the form of the muscles of the Adductor Group.

comparative note

For comparative purposes the Chaffinch, Fringilla coelebs, and the Brambling,
F. montifringilla, have been selected as more generalized fringilline species.

In comparison with C. coccothraustes the oral surfaces of the rhamphotheca of
F. coelebs are smooth and lack any dilations similar to the striated pads, or bosses
(Text-fig. 8b) . It would appear (Sushkin, 1925) that this specialization is not confined
to C. coccothraustes for apart from Hesperiphona of North America the occurrence of
specializations of this nature are found in a few other Old World genera, namely,
Eophona, Perissospiza and Mycerobas.

In comparison with the skull of C. coccothraustes that of F. coelebs is a fragile
structure and this difference is illustrated by comparing the sizes and the weights of
the skulls of the two species. The average maximum dimensions of skulls

386

THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH

of C. coccothraustes are 40 x 20 x 20 mm. and F. coelebs 30 x 14 X 15 mm. ; if
the cube of each of the linear proportions of the skull of C. coccothraustes is compared
with those of a hypothetical skull of the same average dimensions as F. coelebs, but
of the same design as that of C. coccothraustes, it is found that the dimensions of the
skull of C. coccothraustes average 250% " larger ". Yet the average weights of the
skulls of C. coccothraustes and F. coelebs are 3*235 gm. and 0-654 § m - respectively,
that is, the skull of C. coccothraustes contains nearly 400% more bony material than
that of F. coelebs. Therefore, the skull of C. coccothraustes is very massive for its
size.

The outline of the skull of C. coccothraustes has been likened to one horn of a
crescent, but the skull of F. coelebs differs by being more cuneiform, and its dorsal
profile is nearly straight (Text-fig. 8a). The difference is attributable mainly to the

post. orb. proc

for. md.

Fig. 8. Fringilla coelebs. a, Cranium and upper jaw lateral view from the right side
(compare Text-fig. id), b, Lower jaw (compare Text-fig. 2b).

greater angle between the basi-cranial and basi-maxillary axes 1 in F. coelebs which is
150 but only 120 in C. coccothraustes.

Van der Klaauw and Duijm consider that a difference of this nature can be a
functional adaptation and Duijm stated, "... the bill and the cerebral capsule
behave mainly as independent functional elements.". In the present instance the
downward rotation of the bill relative to the cranium in C. coccothraustes reduces
some of the stresses set up by the application of powerful forces to crack fruit stones.
This is shown diagrammatically in fig. 9. Here BC and AB represent the basi-
maxillary axis and a line parallel to the basi-cranial axis in two skulls of the same
overall length. In (a) the subtended angle between the two is 120 as in C. cocco-

1 Basi-maxillary axis: The line of intersection between the vertical longitudinal plane of the head
and the mean ventral surface of the premaxillae.

THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH

387

thraustes, and in (b) it is 150 as in F. coelebs. Assuming that the same forces were
applied to the lower jaw to crack a nut held at the same distance from the angle of
the jaw, then the magnitude and direction of the force, relative to the upper jaw,
can be represented as the line XY. The direction of the force is tangential to a circle
whose centre is at the articulation of the lower jaw, A, and periphery at D, where
the lower mandible presses against the stone. The force XY is resolvable into two

Fig. 9. Diagram showing the main forces imposed on the upper jaw when food is
crushed between the mandibles, a, The subtended angle, ABC, between the basi-
maxillary and basi-cranial axes is 120 . b, The subtended angle is 150 .

components at right angles to one another, XP and XQ. The former exerts a
compressional and the latter a bending strain on the upper jaw. A bending strain
is clearly more likely to fracture the bill than a compressional strain along its length
and the bending component is approximately 6% less in (a) where the angle between
bill and cranium is similar to that of C. coccothraustes. Therefore, the downward
rotation of the bill in C. coccothraustes may be regarded as an adaptation to feeding
so that hard stones are cracked with a smaller bending strain being experienced by
the bill ; that is, there is less chance of fracture than in a skull of the design of
F. coelebs where the subtended angle between the bill and cranium is more obtuse.

388 THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH

The bending strain which is experienced by the bill of C. coccothraustes when fruit
stones are cracked must nevertheless still be relatively enormous when forces in the
region of ioo lb. have to be applied to crack them, but its effect is largely nullified
by the highly ossified septa. These are markedly different in F. coelebs where the
interorbital septum exhibits a more normal passerine condition forming only a thin
bony partition between the orbits and the nasal septum appears to be mainly
unossified.

Again, a high degree of ossification is found in the pterygoid-quadrate-zygoma
system, for in comparison the pterygoids and zygomae of F. coelebs are long and
slender, while the quadrates of C. coccothraustes are about twice the size of those of
F. coelebs. 1

An examination of the skulls reveals that not only has the skull of C. coccothraustes
become modified to withstand greater stress, but also to provide a greater area of
the attachment and insertion of the jaw muscles. This latter point is well illustrated
by the difference observed between the lower jaw of C. coccothraustes and F. coelebs.
In the former the coronoid is elevated in correlation with the large size of the pars
medius of the M. Adductor mandibulae externus which is inserted into it, moreover,
the foramen mandibularis is very small while it is relatively large in F. coelebs
(Text-fig. 10). The disparity in the sizes of the partes medii in the two birds is

Dp.

p.p.

Fig. io. Fringilla montifringilla. Diagrammatic view of the right side of the head
showing the external jaw muscles (compare Text-fig. 3).

also reflected in the degree of development of the respective post-orbital processes
of the squamosals from which they originate. Generally the lower jaw of F. coelebs
appears to be more primitive in character than that of C. coccothraustes which, in
addition to the adaptations just referred to, is modified to carry the heavy bosses of
the bill, as described above.

1 Quadrate of C. coccothraustes : mass 0-0375 g m -» 9 m m. high x 7 mm. long (orb. proc). Quadrate
of F. coelebs : mass 0-0041 gm., 4-5 mm. high x 4-5 mm (orb. proc).

THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH 389

The jaw muscles of C. coccothraustes in comparison with those of the Brambling,
Fringilla montifringilla, are exceedingly tendinous. In the latter they are composed
of fleshy fibres almost devoid of tendons, except at the site of insertion, and even then
the tendons are not as evident as those in C. coccothraustes. The jaw muscles of
F. montifringilla are generally more bulky and appear to contain much non-con-
tractile tissue. One gains the impression that the flaccid muscles of F. montifringilla
are incapable of the rapid powerful contractions of the muscles of C. coccothraustes.

The difference in the form of the M. Adductor mandibulae internus (M. Pseudo-
temporalis) in the two species illustrates the comparatively compact nature of the
muscles of C. coccothraustes. In this bird the M. pseudotemporalis is squarish in
cross-section and extends from the posterior wall of the orbit to the lower jaw. In
F. montifringilla, on the other hand, the M. pseudotemporalis although similarly
attached and inserted forms a fleshy floor to the orbit filling the spaces between the
Ms. Adductor mandibulae posterior, Pterygoideus dorsalis and Ethmo-mandibularis.
Yet, despite its size, the muscle probably does not contract with the same forces as its
smaller counterpart in C. coccothraustes.

The nature of the jaw muscles in C. coccothraustes is not the only factor contributing
to a powerful musculature, but their size and distribution are also important. A
comparison of the relative sizes and the areas of attachment of the various portions
of the M. Adductor mandibular externus illustrates this point. It can be seen
(Text-figs. 3 and 10) that the area of attachment of this muscle in C. coccothraustes
is disproportionately greater than in F. montifringila. In C. coccothraustes the
muscle arises from over most of the external surface of the cranium whereas in
F. montifringilla the areas of attachment are confined to the lateral and postero-
lateral surfaces.

The essential difference between the myology of the jaw of C. coccothraustes and
that of F. montifringilla appears to be one of degree, that is, area of attachment and
tendinosity. However, what I have termed the " anterior portion " of the M.
Pterygoideus ventralis lateralis appears to be absent in F. montifringilla. In this
species the M. Ethmo-mandibularis is more prominent and its site of origin extends
over the comparable area of origin of the anterior portion of the M. Pterygoideus
ventralis lateralis in C. coccothraustes. It would appear from this that the M. Ethmo-
mandibularis is divided into two parts in C. coccothraustes. However, since in C.
coccothraustes the anterior portion is inserted into the palatine along with the posterior
portion of the M. Pterygoideus ventralis lateralis I am of the opinion that the anterior
portion should be considered as a part of the M. Pterygoideus ventralis lateralis,
and this seems justifiable since the portions are also functionally complementary.

DISCUSSION

The purpose of this paper has been to indicate some of the modifications in the
morphology of the head of C. coccothraustes which appear to be related to its ability
to crack open various fruit stones, such as those of the cherry, damson and olive.
There are, however, several points requiring further investigation.

39o THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH

There is, for example, a point of interest in the development of the horny pads of
the rhamphotheca. It seems that these pads, or bosses, do not appear until the bird
reaches maturity. In the limited sample examined they were not found in nestlings
nor in birds in immature plumage. Even if pads were present in young birds they
would be non-functional because from an examination of the stomach contents it
appears that the adults are primarily insectivorous during the breeding season
(Mountfort, in press) which indicates that the nestlings are fed mainly on insects.
Similarly, the pads are not required by immature birds because they have been ob-
served to feed on caterpillars, especially the Hornbeam caterpillar, Carpinus betulus.

The appearance of the horny pads late in the development of the individual could
be interpreted as evidence that the structures have been acquired recently in the
history of the species. However attractive this argument may appear to be at first
sight it is misleading because the pads should not be considered apart from the asso-
ciated osteological and myological modifications ; and it is possible that if the pads
were present in an immature bird they would even be a danger to it ! One of the
chief characteristics of young passerine birds is the absence of osseous material in
most of the cranium and the slow rate at which ossification occurs as the young bird
matures. The absence of osseous material makes the skull extremely fragile, the
more so since the sutures between the bones remain open almost until maturity.
The skull of a young bird is, therefore, not strong enough either to wishstand the
stresses of cracking open fruit-stone kernals or to accomodate the powerful muscu-
lature capable of closing the jaws with a force in the region of ioo pounds, that is if
the muscles could be precociously developed to contract with that force. It seems,
therefore, that the late development of the pads cannot be regarded as indicative
of their appearance in the phylogeny of the bird.

CONCLUSIONS

The structural modifications which have occurred in the head of C. coccothraustes
and enable the bird to apply and withstand a force in the region of ioo lb. are pro-
found. It appears that no part of the head, which has been considered, has escaped
specialization. The oral surfaces of the horny bill are modified and equipped with
striated pads between which the food is gripped. The skull, particularly the upper
and lower mandibles, is strengthened to withstand the stress of cracking a hard fruit
stone and it is modified to accommodate the powerful musculature of the jaw as well
as to withstand the force of the contraction of the muscles. The musculature of the
head is highly developed with large areas of attachment, and the individual muscles
are tendinous in nature and powerful in action.

The possession of these modifications are undoubtedly a selective advantage to
the bird particularly during the period of great increases in population which partly
coincides with the " soft-fruit " season. C. coccothraustes comes less into competition
with other seed-eaters and its own young at an important time of the year by utilizing
fruit-stones and similar large seeds, berries and, during the breeding season, large
insects.

THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH

391

APPENDIX

Crushing Tests on Cherry and Olive Stones

(a) The National Physical Laboratory carried out a series of crushing tests and made
the following report :

The apparatus used for the tests was designed to simulate, as far as possible, the
mandibles and pressure pads of this bird. It consisted of a steel compression rig
with two J-in. diameter rods which fitted into two deep grooves in the lower block
of the apparatus, and a flat serrated cross piece attached to the upper block. The
bare fruit stone was placed between these loading points and cracked ; the load
being increased by moving a jockey weight along the lever of the testing machine.
The time taken to crack each stone was approximately 20 seconds.

In order to represent the influence of the direction of the seam or suture of the
stone in the bird's beak, the tests were made with the seam facing several directions.

The results of the tests are given in Tables I and II.

(b) Mr. D. B. Welbourne, of the Department of Engineering, Cambridge, working
independently obtained comparable values when he crushed cherry stones in a
Housefield Tensometer. In correspondence, he writes, that in his tests he effected
different rates of loading and found that more rapid loading resulted in higher failure
loads, that is, a greater pressure is required to break a stone quickly than to break it
slowly. Therefore, the crushing loads given in Tables I and II should not be regarded
as the maximum pressures applied by C. coccothraustes when cracking fruit stones ;
for the values, determined by the tests, were obtained in each case by applying the
pressure for approximately 20 sec. which is a greater time than it takes C. cocco-
thraustes to crack a fruit stone.

Test

1
2
3
4
5
6

7
8

9
10

2
3

Table I. — Crushing

tests on

Cherry Stones

Direction of

Type of Cherry

Crushing load

seam

(lb.)

Sideways

Dark

62
70
62

Upwards

>t
>>

70
65

Downwards

»>

60
62

>»

Sideways
>>

Light

95
93

Upwards
»

70
68
70

392

THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH

Table II. — Crushing tests on fresh Olive Stones

"est

Direction of Type c

)f stone

Crushing load

seam-

(lb.)

Sideways . Olive

>> •

106

114

Upwards

125

„

137

147

143

Downwards

136

107

138

>> •

159

Physic's Division.
National Physical Laboratory.

REFERENCES

Beecher, W. J. 1953. A phylogeny of the Oscines. Auk, 70 : 270-333.

Duijm, M. 1 95 1. On the head posture in birds, and its relation to some anatomical features.

Pt. I and II. Proc. Kon. Ned. Akad. Wentensch. Ser. C, 54 : 202-211 ; 260-271.
De Beer, G. R. 1951. Embryos and Ancestors. Oxford.
Edgeworth, F. H. 1907. The development of the Head-muscles in Gallus domesticus and

the Morphology of the Head-muscles in the Sauropsida. Quart. Jour. Micro. Sci. N.S. 51 :

511-556.

1935- The Cranial Muscles of Vertebrates. Cambridge.

Fiedler, W. 1951. Beitrage zur Morphologie der Kiefermuskulatur der Oscines. Zool.

Jahrb. (AnaL), Jena, 71: 235-288.
Gadow, H. 1 891. Vogel : Aves. in Bronn's Klass n. Ord. des Thier-Reichs. 6.
Hale, J. R. 1945. Great-spotted Woodpecker breaking open almond nuts. Brit. Bds.

38 : 17.
Hofer, H. 1945. Untersuchungen iiber den Bau des Vogelschadels, besonders iiber den der

Spechte und Steisshuhner. Zool. Jahrb. (Anat.) Jena, 69 : 1-158.

1949. Die Gaumenlucken der Vogel. Acta. Zool. Stockholm, 30 : 209-248.

1950. Zur Morphologie der Kiefermuskulatur der Vogel. Zool. Jahrb. [Anat.) Jena,

70 : 427-556.

Ingram, C. 1941. Hawfinches and Cherries. Brit. Bds. 34 : 220.

James, T. O. 1945. Great-spotted Woodpecker breaking open almond nuts. Brit. Bds.

38 : 274-275.
Kirkman, F. B. 1911. The British Bird Book, 1 : 155. T. C. & E. C. Jack, London.
Klaau w, C. J . , van der . 1 948-1 952. The size and the position of the functional components of

the skull. A contribution to the knowledge of the architecture of the skull. Extrait des

Arch. Neerl. Zool. 9 : 1-4.
Kripp, D. von. 19330. Der Oberschnabel-Mechanismus der Vogel. (Nach den Methoden

der graphischen Statik bearbeitet). Gegenbaur, Morph. Jahrb., Leipzig, 71 : 469-554.

I933&- Der Spezialisationsreiher der Storche, Reiher und Kormorane von konstruktiven

und biotechnischen Standpunkt. Gegenbaur. Morph. Jahrb., Leipzig, 72 : 60-92.

I933C Beitrage zur mechanischen Analyse des Schnabel-Mechanismus. Gegenbaur,

Morph. Jahrb., Leipzig, 72 : 541-556.

THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH 393

Lakjer, T. 1926. Studien uber die Trigeminus-versorgte Kaumuskulatur der Sauropsiden.

Kopenhagen. C. A. Reitzelg.
Lebedinsky, J. 1 92 1. Zur Syndesmology der Vogel. Anat. Anz. Jena, 54 : 8-15.
Lebedinsky, N. G. von. 1921. Der Unterkiefer der Vogel. Acta. Univers. Latviensis, 1 : 12-

44.
Moller, W. 1930-1931. Uber die Schnabel- und Zungenmechanik bliitenbesuchender Vogel.

I und II. Ein Beitrag zur Biologie des Blumenvogels. Biologia Generalis. 6 : 651-726 ;

7 : 99-154-
Mountfort, G. R. The Hawfinch (in press). London, Collins.
Owen, R. 1866. The Anatomy of Vertebrates. London, Longmans Green & Co.
Parker, W. K. 1877-1879. On Aegithognathous Birds. Trans. Zool. Soc. Lond. 9 : 289-352;

10 : 251-314.
Pycraft, W. P. 1905. Bull. Brit. Ornith. Club, 15 : 37.
Sharpe, R. B. 1888. Cat. Bds. Brit. Mus. 12 : 1-3.
Shufeldt, R. W. 1890. The Myology of the Raven, MacMillan, London.
Sushkin, P. 1925. The Evening Grosbeak (Hesperiphona), the only American genus of a

Palaearctic group. Auk, 42 : 256-261.
Streseman, E. 1927-1934. Aves. In Handb. d. Zool. herausgegeb. von Th. Krumbach,

7: 2.
Watson, M. 1883. Report on the anatomy of the Spehniscidae collected. 'Challenger'

Report. Sci. Results. Challenger, Zool. 7, H.M.S.O., London.
Witherby, H. F., et al. 1939. The Handbook of British Birds, 1 : 51-54, Witherby, London.
Witherby, H. F. & Ticehurst, N. F. 1907. On the more important additions to our know-
ledge of British Birds since 1899. Brit. Bds. 1 : 152.

PRESENTED
1955

ZOOL. 2, 13.

INDEX TO VOL. II

The page numbers of the principal references are printed in Clarendon type.
New taxonomic names are printed in bold type.

abyssicola, Ophiactis

140, 154

bartletti, Tretocidaris

140

Acanthephyra

123, 125

Bathothauma

• 125

Acanthocyclus

Bathypolypus

. 97

Acartia

133

beatrix, Aphrocallistes .

215, 218

Achirus

134

Bentheledone

. 97

ackleyi, Zoroaster .

139

Berlesiana

. 295

aculeata, Ophiopholis

140

bidentata, Ophiacantha .

140

adspersum, Ophiernus

140

biguttatus, Portumnus .

62, 63

adspersus, Ophiernus

159

borealis, Lissodelphis

• 343

africana, Ilisha

137, 163, 211

Bosmina

• 136

afzeliusi, Pellonula

137. 163, 211

botryoides, Leucosolenia

215, 218

alaskensis, Haemogamasi

is 287

bower banki, Halichondria

. 216, 233

albula, Stephanasterias

• i39. 151

bracebridgei Crepidacantha

• 255

alpinus, Salvelinus

163, 183

brevicaudum, Ophioderma

140

amaranthus, Phorbas

216, 228

Brisingella

• 139, 151

Amphiophiura

140, 158

burgeri, Panulirus .

25, 35

Amphipholis .

140, 155

byronia, Otaria

. 33i» 332

Amphiura

140

anancora, Stelletta

. 216, 220

anhelans, Tedania .

. 216, 239

caecus, Phoberus .

. 124

annexa, Desmacella

216, 226

calcinea, Haliclona

. 216, 224

annulata, Astroporpo

140

californianus, Zalophus .

■ 334

Anthenoides .

• 139

Callinectes

• 137

antillaru m , Plesiodiadem

1 140

Calliteuthis .

. 125

antillensis, Astropecten

. 139

Callorhinus .

• 334

Aphrocallistes

124, 215, 218

Callyspongia .

. 216, 225

Arctacaridae .

287-307

Calveriosoma .

130, 140

Arctacarus

. 299

canariensis, Leucosolenia

215, 219

Arctocephalus

311-336

Cancer .

60, 64

Arctoseius

. 287

Capromys

. 124

arctuatus, Zercon .

. 296

Carcharinus .

125, 163, 164

arenatus, Nymphaster

. 139

Carcinides

• 59

argentata, Otaria .

• 323

carinata, Tropiometra

. 140

argus, Puerulus

17. 33

carpenteri, Pheronema .

215, 217

articulatus, Astropecten

• 139

catodon, Physeter .

• 103

assimilis, Hymeniacidon

. 216, 231

Caulerpa

. 130

Astropecten .

. 139

Ceramaster .

139. 148

Astrophyton .

124,140

cetratus, Arctoseius

. 289

Astroporpa .

140

Chaetodipterus

• 133

Atelecyclus .

Chiridota

140

aurantiifrons, Ptilinopus

272, 275

Chrysoena

. 278

australis, Arctocephalus

311, 324, 333

cincta, Ptilinopus .

. 272, 274

Axinella

216, 229

clathrata, Luidia .

• 139

clavicorne, Phyllosoma .

25, 29

coccothraustes, Coccothraustes

371-391

balssi, Trachycarcinus

• 59

coelebs, Fringilla .

• 372

balteatus, Ceramaster grs

inularis 139, 148

Coelorhynchus

. 130

bandensis, Octopus

Colomesus

• 133

barimae, Luidia

• 13

9, 145

Columba

. 279

396

INDEX

corallinus, Trachycarcinus 59

fischeri, Ptilinopus

272, 274

coregoni, Bosmina .

. 136

fluviatilis, Labidocera

• 133

cornuloides, Halichondria

216,234

Fringilla

• 372

coronata, Brisingella

139, 151

Fungia

• 123

coronulatus, Ptilinopus

272, 276

Corystes

galapagoensis, Arctocephalus

. 311

Cosinodiscus .

• 135

galilaea, Tilapia

• 35i

cranium, Tetilla

216, 223

gargantua, Mesothuria

140

Crepidacantha

243-261

giesbrechti, Acartia

• 133

crinispina, Crepidacantha

. 248

girigan, Tilapia

• 35i. 355

crispatus, Ctenodiscus

. 139

glabrata, Hymeniacidon .

. 216, 231

Crossaster

. 139

glaucus, Trachycarcinus .

47. 59

Ctenodiscus . . .

. 139

Globicephala

104

curiosus, Zercon

296

goliath, Ducula

270, 280

cuspidatus, Goniaster

• 139

Goniaster

• 139

Cystoseira

44. 49

gracilis, Octopus

Cystisoma

• 125

gracillima, Amphipolis .

140, 155

Cythere

granadensis, Ophioplinthaca

140, 151

grandis, Crepidacantha .

. 252

granulans, Ceramaster .

139, 148

danae, Macrotritopus . . . 70-97

granulifrons, Ptilinopus .

• 273, 277

dasypus, Panulirus . . 17, 25, 31, 35

greyii, Ptilinopus .

272, 276

denticulata, Perimela . . -57, 60-63

grimaldii, Spherosoma

130, 140

Desmacella ....

216, 226

guttatus, Panulirus

diadema, Stephanotrochus

. 123

Didunculus ....

. 280

Haemogamasus

. 287

Dinychus ....

• 305

Halichondria

. 216, 232

Diplobatis ....

. 163, 166

Haliclona

. 216, 223

distorta, Myxilla

216, 226

Halimeda

. 130

dohertyi, Ptilinopus

272, 274

Haplochromis

• 35i

dolabriformis, Ophionereis

140, 155

Harengula

. 163, 211

Dragmatyle ....

216, 235

Henricia

• 139

Drepanoptila

. 278

Hesperiphona

• 372. 385

drobachiensis, Strongylocentrotus

140

Higginsia

. 2i6,235

Ducula ....

270, 278

Hircinia

. 216, 236

dupetithouarsii, Ptilinopus

273, 276

Holothuria .

140

homarus, Panulirus

17-41

hoylei, Calliteuthis

. 125

Echinaster . . . . . .139

hunteri, Tilapia

• 364

echinata, Ophiocoma

140

huttoni, Ptilinopus

269, 273, 275, 276

elegans, Octopus

Hymeniacidon

. 216, 231

elegans, Ophiolepis

140

hyogastra, Ptilinopus

• 273, 277

elephas, Palinurus .

hystrix, Calveriosoma

130, 140

elongatus, Johnius

• 137

Eophona

• 385

Ilisha ....

137. 163. 211

Epicrioidea

• 295

Inachus .

Epicrius

. 294

Incisa, Ophioplinthaca .

140

equivocus, Octopus

.70.71

inermis, Puerulus .

Erimacrus

infernalis, Vampyroteuthis

. 124

erina, Haliclona

. 216, 223

Inflatella

. 216, 228

esculenta, Tilapia .

• 35L 364

inflatus, Panulirus .

• 28, 33, 34

Eumetopias .

• 332

inornata, Desmacella

216, 226

excentricus, Coscinodiscus

• 135

insolitus, Ptilinopus

• 273, 277

insularis, Ptilinopus

273, 276

iozonus, Ptilinopus

• 273, 277

faber, Chaetodipterus . . . 133

facilis, Pellisolus

206

Farrea

215, 217

jambu, Ptilinopus

273, 278

fenestralis, Zercon .

. 297

japonicus, Panulirus

17. 3i

fenestrata, Thenea

215, 219

Jasus .

. 17, 18, 20

INDEX

397

jipe, Tilapia .

Johnius

jubatus, Eumetopias

Justitia

kempi, Macrotritopus
kirkpatricki, Crepidacantha
koefoedi, Searsia
korogwe, Tilapia mossambica

kroyeri, Xiphopenaeus .

labiatus, Coelorhynchus
Labidocera .
Laelaptidae .
laevis, Chiridota
lalandei, Jasus
laterincisus, Actoseius
latimanus, Callinectes
latipes, Portumnus
layardi, Ptilinopus
leclancheri, Ptilinopus
Leptasterias .
Leptoderma .
leucas, Carcharhinus
Leucosolenia .
Leucotreron .
lilljeborgi, Acartia .
Lissodelphis .
longimana, Justitia
longimanus, Carcharhinus
longiseta, Crepidacantha
Luidia

luteovirens, Ptilinopus .
lutzi, Pseudoceradocus .
lyromma, Bathothauma .
lytteltonensis, Traehyleberis

macrochir, Tilapia .
Macropodia
macrops, Leptoderma
Macrotritopus
maenas, Carcinus .
marchei, Ptilinopus
marginatus, Astropecten
Megaloprepia
melaena, Globicephala
melanopleura, Tilapia
melanospila, Ptilinopus
meleagris, Stomolophus
mercieri, Ptilinopus
merrilli, Ptilinopus
Mesothuria .
metabula, Amphiophiura
metallacta, Ophiacantha
mexicana, Holothuria
micropunctatus, Dinychus
monacha, Ptilinopus

351, 354

• 137

• 332
17, 18

70-97

. 256

[63, 201, 206

• 351, 356

• i34

130

133

287

140

288

137

62, 63

273, 278

272, 274

• 139
163, 183
163, 164
215, 218

269, 272, 273, 278
i33» 135
341-345
18
164

• 251
139, 142

273, 278

• 134

• 125

• 3-i5

• 35i
59

163, 183
69-97

• 58

272, 274
. 139
. 279

104

• 35i

273. 277
33. 134. 135

273, 276

272, 274

140

140, 158

140

• 305
272, 276

Monodi, Sirpus
montifringilla, Fringilla
mortenseni, Rossella
mossambica, Tilapia
Mugil .

multidentatus, Arctoseius
muricatum, Astrophyton
Mycerobas
Myxilla

naina, Ptilinopus .
natalensis, Tilapia .
nigra, Tilapia
nilotica, Tilapia
nodosa, Stegophiura
nodulosus, Xenodermichthys
Normichthys
nuchalis, Tachysurus
Nymphaster

occa, Farrea

obscurus, Carcharinus

occipitalis, Ptilinopus

Octopus

odontostoma, Crepidacantha

operosa, Normichthys

Ophiacantha

Ophiactis

Ophiernus

Ophiocoma

Ophiocten

Ophioderma

Ophiolepis

Ophiomitra

Ophiomusium

Ophionereis

Ophiopholis

Ophioplinthaca

Ophiura

ornatus, Arctoseius

ornatus, Panulirus

ornatus, Ptilinopus

osculum, Halichondria

Otaria .

Otariidae

pachastrelloides, Stryphnus

pagurus, Cancer

Palaemon

panicea, Halichondria

papposus, Crossaster

Palinuridae .

pangani, Tilapia

Panulirus

Parazercon

parvipora, Crepidacantha

Pellisolus

Pellonula

crinispina

44, 52-56

• 372

215, 218

• 35i

• 134
289, 291
124, 140

• 385

216, 226

273. 277
358
35i
35i
140
196
206
134
139

215, 217
. 164

272, 274, 281

69-97

. 258

206

140

140, 154

140, 159

140

140, 159

140, 155

140

140, 151

140

. 289

• 28, 31, 41

272, 275, 281

216, 233

• 323
311-336

216, 220

• 64

• i34
216, 232

• 139
17-42

35i. 356
17-42

• 295
250
206

i37» l6 3. 211

398

INDEX

pellucidus, Phyllophorus

pellucidus, Puerulus

Peltarion

Penaeus

penicillatus, Panulirus

pensacolae, Harengula

pentacrinus, Ophicantha

Perissospiza .

perlatus, Ptilinopus

peroni, Lissodelphis

perousii, Ptilniopus

Persparsia

Pheronema .

philippi, Arctocephalus

Phoberus

Phocaena

phocoena, Phocaena

Phorbas

Phormosoma

Phronimae

Phyllophorus

Physeter

Phytoseiidae

pictus, Diplobatis .

piercei, Anthenoides

pinnatus, Synaphobranchus

Pirimela

placenta, Phormosoma

Plesiodiadema

Plocamionida

Podocactes .

poissonii, Crepidacantha

polaris, Leptasterias

polycoeca, Searsia

polyphagus, Panulirus

porphyraceus, Ptilinopus

porphyrea, Ptilinopus

Portunus

Prodinychidae

Prozecon

Pseudoceradocus .

Pseudocorystes

psittacus, Colomesus

Ptilinopus

Puerulus

pulchellus, Ptilinopus

purpuratus, Ptilinopus

Radiella

ramosa, Axinella
Ramphiculus
rarispina, Salenocidaris
rarotongensis, Ptilinopus
regina, Ptilinopus .
regius, Panulirus
richardsii, Ptilinopus
rivoli, Ptilinopus .
Rosaura

140

133. 134

163, 211
140

• 385
272, 275

341-345

272, 276

63, 206 ,208

215, 217
11, 322, 333

. 124

. 104

104

216, 228
140

• 125
140

• 103
. 287

163, 166
. 139

• 130
44

. 140

140

216, 229

244, 246

• 139
201, 206

28, 31, 41

272, 276

70, 272, 274

• 305

• 295

• 134
60

133. 134
267-284

17
272, 276
270, 273, 276, 281

. 216, 222

216, 229

269, 272, 274, 278

140

273, 276

272, 276

271, 273, 277

. 163, 167

rosaurae, Luidia

rosaurae, Sphaerodactylus

roseicapilla, Ptilinopus

Rossella

rostratus, Arctacarus

rostratus, Searsia

rotunda, Rosaura .

rubens, Haliclona .

Sagitta

Salenocidaris

Salvelinus

sanguinolenta, Henricia

sarsi, Ophiura

savignyi, Ophiactis

scabrocuneata, Cythere

scabrocuneata, Trachyleberis

Scaeurgus

scorpio, Macrotritopus

Searsia .

Seiodes

senegalensis, Luidia

sentus, Echinaster

sericeum, Ophiocten

setigera, Crepidacantha

Sirpus .

sol, Radiella .

solea, Crepidacantha

solenites, Zercon

solomonensis, Ptilinopus .

Sperosoma

Sphaerodactylus

spiculosa, Haliclona

spinnulifer, Trachyc;

Spirula

squinado, Maia

Stegophiura .

Stelletta

Stellifer

stellifer, Stellifer

Stenella

Stephanasterias

Stephanotrochu s

Stomolophus

strigilata, Higginsia

Strongylocentrotus

Stryphnus

subgularis, Ptilinopus^

sublaevis, Dinychus

sundevalli, Amphiura

superbus, Ptilinopus

Symphurus .

Synaphobranchus .

taaningi, Persparsia
Tachysurus .
tannensis, Ptilinopus
Tedania
Telmessus

163.

139,

142

123

272,

276

. 215,

218

299-305

206

. 163,

167

216,

223

• 133.

135

140

. 163,

183

139

140

4. II

74-97

70-97

199, 201,

206

295

139

140

250

43-64

216,

222

252

296

271. 273,

277

• 130.

140

123

216,

224

122

140

216,

220

134

124

• 139,

151

123

133

216,

235

140

216,

220

272,

274

306

140

272,

276

i34

130

163, 206,

208

134

270, 272,

275

216,

229

tenera, Halichondria

tenerrima, Callyspongia .

tenerrima, Haliclona

teres, Crepidacantha poissonii

Tethyaster

Tetilla

tetracirrhus, Pteroctopus

Thenea

Thia ....

thoracatus, Capromys

Tilapia

topsenti, Dragmatyle

topsenti, Plocamionida

townsendi, Arctocephalus

Trachycarcinus

Trachyleberis

Treron ....

Tretocidaris .

triangularis, Zercon

Trichopeltarion

Tripneustes

Tropiometra .

unicirrhus, Scaeurgus
ursinus, Callorhinus
ursinus, Seiodes

valida, Ophiomitra
validum, Ophiomusium

INDEX

• 232

Vampyroteuthis

216, 225

variabilis, Hircinia

216, 225

variabilis, Tilapia .

• 247

ventricosus, Tripneustes

139, 141

verrucosa, Maia

216, 223

versicolor, Panulirus

96, 97

vesitus, Tethyaster

215, 219

vespertina, Hesperiphona

victor, Ptilinopus .

. 124

viridis, Innatella .

349-368

viridis, Ptilinopus .

216, 235

vulgaris, Octopus .

216, 229

vulgaris, Palinurus

. 321

44-64

. 3-15

wallacii, Ptilinopus

■ 279

weberi, Arctoseius

140

wollebaeki, Zalophus

296

wy villi, Thenea

140

Xenodermichthys .

Xiphopenaeus

70-97

• 334

Zalophus

• 295

zariquieyi, Sirpus .

zelanica, Crepidacantha .

Zercon .

140

Zerconoidea .

140, 159

Zoroaster

399

. 124
216, 236

35L 364
140

58,59

25, 26, 31

i39, 141

• 372

273, 278

216, 228

273, 277

. 70

272, 275

289, 291

. 320

215, 220

196
134

320, 334
•43-5i

• 254

294, 296

295, 296

• 139

1 2 FEB \m

PRINTED IN GREAT BRITAIN BY
ADLARD AND SON, LIMITED,
BARTHOLOMEW PRESS, DORKING.

■